A recent conversation on Twitter led to the suggestion that someone compile a list of physiological effects of concern for BOLD. That is, a list of potentially confounding physiological changes that could arise sympathetically in an fMRI experiment, such as altered heart rate due to the stress of a task, or that could exist as a systematic difference between groups. What follows is the result of a PubMed literature search (mostly just the abstracts) where I have tried to identify either recent review articles or original research that can be used as starting points for learning more about candidate effects. Hopefully you can then determine whether a particular factor might be of concern for your experiment.

This is definitely not a comprehensive list of all literature pertaining to all potential physiological confounds in fMRI, and I apologize if your very important contribution didn't make it into the post. Also, please note that I am not a physiologist so if I go seriously off piste in interpreting the literature, please forgive me and then correct my course. I would like to hear from you (comments below, or via Twitter) if I have omitted critical references or effects from the list, or if I have misinterpreted something. As far as possible I've tried to restrict the review to work in humans unless there was nothing appropriate, in which case I've included some animal studies if I think they are directly relevant. I'll try to keep this post up-to-date as new studies come out and as people let me know about papers I've missed. As it says at the top, I'll consider this a draft post pending feedback. Subsequent posts will be designated with a version number.

A final caution before we begin. It occurs to me that some people will take this list as (further) proof that all fMRI experiments are hopelessly flawed and will use it as ammunition. At the other extreme there will be people who see this list as baseless scare-mongering. How you use the list is entirely up to you, but my intent is to provide cautious fMRI scientists with a mechanism to (re)consider potential physiologic confounds in their experiments, and perhaps stimulate the collection of parallel data that might add power to those experiments.

Getting into BOLD physiology

There are some good recent articles that introduce the physiological artifacts of prime concern. Tom Liu has reviewed neurovascular factors in resting-state functional MRI and shows how detectable BOLD signals arise from physiological changes in the first place. Kevin Murphy et al. then review some of the most common confounds in resting-state fMRI and cover a few ways these spurious signal changes can be characterized and even removed from data. Finally, Dan Handwerker et al. consider some of the factors causing hemodynamic variations within and, in particular, between subjects

Once you start really looking into this stuff it can be hard not to get despondent. Think of the large number of potential manipulations as opportunities, not obstacles! Perhaps let The Magnetic Fields get you in the mood with their song, "I don't like your (vascular) tone." Then read on. It's a long list.

Breathing and heart rates

These are undoubtedly the two big issues and will be top-of-mind for most people. Murphy et al. do a good job introducing the topic of cleaning raw data, e.g. using independent respiration and heart rate data. There are numerous correction methods available today (see the section, "Time series cleanup with physiological recordings" in the Murphy paper) and these references are perhaps the best places to start if you want to dig into the origins of the problems. In reverse chronological order we have:

PHYCAA+ by Churchill & Strother
GLMdenoise by Kay, Rokem, Winawer, Docherty & Wandell.
RIPTiDe by Frederick, Nickerson & Tong (using near-infrared spectroscopy, NIRS)
RVT-HR by Chang, Cunningham & Glover
RVT by Birn, Diamond, Smith & Bandettini
RETROICOR by Glover, Li & Reese

Then there are also numerous ICA-based methods, including multi-echo acquisition with a model fit. I won't comment on the relative or absolute performance of any of these schemes, just let you know you have options! But one important lesson seems to be this: measuring heart rate and respiration independently isn't likely to hurt.

I'll also mention the inter-relatedness of HR and breathing, through vagal tone. This has implications for fMRI studies employing emotional stimuli.

CO2: hypercapnia and hypocapnia

Inspired CO2 is a well-known and powerful vasodilator. Similarly, breath holding is known to lead to hypercapnia (increases in arterial CO2) and thus vasodilation. Thus, under normal conditions (no external CO2 source) the arterial concentration of CO2 will be closely linked to the breathing rate, as demonstrated in this 2004 study. The quantitative relationship between breathing rate and end-tidal CO2 is further investigated in this 2009 study.

As for dealing with the variation, in a 2012 study it was demonstrated that clamping arterial CO2 to a narrow range improved retrospective physiological artifact removal from resting-state and task-based data. I like this idea because I've often wondered whether not having the magnet bore properly ventilated could cause hypercapnia. I can't find any papers that have looked at this issue, however. The closest I've found is a study that used CO2 supplied artificially, clearly demonstrating the potential interaction of inspired CO2 with stimulus-driven BOLD. (There was an earlier study in rats that reached essentially the same conclusion.) It's probably best to keep the ventilation constant in the magnet bore for all subjects, just in case.

It turns out there may be gender bias, too. A 2010 study found that men have higher cerebrovascular reactivity to a CO2 challenge. So, not only do men tend to exhibit more head movement than women on average, men also appear to be more sensitive to CO2.

O2: hyperoxia and hypoxia

While refluxed CO2 may be a legitimate concern, incidental hyperoxia seems unlikely unless your MRI facility has a very strange design indeed. But, if you happen to be doing fMRI of divers, high-altitude pilots or astronauts in a simulated environment, do bear in mind the possibility of arterial BOLD from dissolved O2 if your subjects are breathing 100% oxygen. I suppose this ought to be a concern in anesthetized fMRI subjects, too. (I'll consider the physiologic consequences of anesthetics below.) If you're comparing normal volunteers (awake) to anesthetized patients then you might want to consider giving the former group the same gas mixture (sans anesthetic) as the latter.

What about hypoxia? If you have a well-designed MRI facility then ordinarily acute hypoxia shouldn't be an issue. But if your scanner is in a poorly ventilated basement you might want to consider an oxygen sensor, assuming you don't already have one, or improving your air conditioning. Otherwise you may be contributing to the variability in your subjects' BOLD responses.

Chronic hypoxia may be an important issue to track, too. Yan et al. studied immigrants who had grown up at high altitude (HA, 2616–4200 m) and relocated to sea level (SL, < 400 m) to a group of sea-level controls, and found both anatomical and resting state fMRI differences associated with "chronic high altitude hypoxia." These changes were in spite of "...no significant differences in hemoglobin levels, circulating red blood cell count, blood pressure, and pulse rate between HA residents and SL controls." In a separate study, Yan et al. found persistent, correlated alterations in verbal working memory and BOLD responses for subjects who had relocated from HA (2,982.8 ± 478.7 m) to SL (< 400 m), even though they had been resident at sea level for at least one year and in spite of consistent hemoglobin levels.

Vasomotion

According to that great oracle, Wikipedia, "vasomotion is the spontaneous oscillation in tone of blood vessels, independent of heart beat, innervation or respiration." Fair enough. Our concern is whether vasomotion might differ across groups, or perhaps might even vary for an individual under different conditions. Again according to Wikipedia, "vasomotion has been shown to be altered in a variety of pathological situations, with vessels from both hypertensive and diabetic patients displaying altered flow patterns as compared to normotensive vessels."

Murphy et al. say this: "If vasomotion is independent of cardiac, respiratory, arterial CO2 concentration and blood pressure fluctuations, its low-frequency characteristics will present another confound for resting-state fMRI BOLD oscillations."

How serious a concern might spontaneous low-frequency variations be for your average BOLD study? A recent paper by Tong & Frederick claims that low frequency oscillations (LFOs) are a larger component of physiologic noise than respiration or heart rate, even when respiration and heart rate frequencies are aliased because of the relatively long TRs that are typical of most fMRI studies. There is also evidence from intraoperative recordings that vasomotion (at ~0.1 Hz) is "spatially localized to distinct regions of the cortex, exhibited wave-like propagation, and involved oscillations in the diameter of specific pial arterioles, indicating that the effect was not the result of systemic blood pressure oscillations."Not small potatoes, then.

What do these LFOs tell us, and (how) do they relate to the ongoing intrinsic neural activity that we assume is driving resting-state fMRI? Methinks this might become the next fashionable study area, if it's not already.

Blood pressure

A good introduction to blood pressure (BP) and its relationship to CBF is given in the Murphy, Birn & Bandettini review. We might be interested in the subject's baseline (pre-scan) BP, but we might also be interested in the BP during a time series acquisition given that BP changes when laying supine versus standing, during stress, etc.

I've been unable to locate any papers showing whether TR-to-TR (real-time) BP is of any use in explaining variance in BOLD time series data. Baseline BP (before MRI) was found to offer only a small normalizing effect on visual-evoked BOLD signals when tested across two conditions; baseline venous T2 was considerably better. A study by Lui et al. in cocaine-dependent (human) subjects found that dobutamine infusion raised mean arterial BP (MAP) but produced only localized BOLD signal changes (in anterior cingulate) that correlated with the BP rise. Gianaros et al. observed a correlation between mean arterial BP and BOLD activity in several brain regions of participants conducting a stressful Stroop task, with BP measured once for each block of sixteen 90-second task blocks.

This is Murphy et al.'s take:

"Evidence of the influence of blood pressure oscillations on resting-state fMRI fluctuations in humans is sparse. Blood pressure levels in rats have been shown to affect evoked fMRI responses, with transient hypertension increasing BOLD (Wang et al., 2006) and CBF (Qiao et al., 2007) signals. Under hypotension, neural activity-evoked CBV increases in visual cortex are negligible compared to ~ 10% at normal blood pressure levels (Nagaoka et al., 2006). Increases in the amplitude of low-frequency BOLD fluctuations have been demonstrated with a drop in mean arterial pressure (Biswal and Kannurpatti, 2009). As supporting evidence in humans, BOLD signal correlates of heart rate and pulse height in the low frequency range have been discovered with fluctuations in cardiac rate explaining up to 11% of the variance in the resting-state BOLD signal (Chang et al., 2009, de Munck et al., 2008 and Shmueli et al., 2007)."

What about chronic hypertension? Another study from Gianaros et al. used the Stroop test to stress healthy young volunteers as a model for determining whether reactivity to psychological stressors might be used to assess risk from hypertension. They found: "Individuals exhibiting greater stressor-evoked MAP reactivity showed (1) greater amygdala activation, (2) lower amygdala gray matter volume, and (3) stronger positive functional connectivity between the amygdala and perigenual anterior cingulate cortex and brainstem pons."

Hypotension may be a factor in cases of major blood loss, perhaps including recent blood donation. While I wasn't able to find any papers looking at recent blood loss (or donation) on BOLD fMRI, Kalisch et al. used a hemorrhage model in rats and observed heterogeneous correlations between BOLD and BP. They suggest that "...a BOLD decrease during a decrease in BP may result from an increase in mostly venous cerebral blood volume (CBV) as an autoregulatory response to maintain cerebral blood flow (CBF) during decreased perfusion pressure."

Hematocrit level

Hematocrit level is the percentage of red blood cells in the blood. And since red blood cells carry hemoglobin, which itself transports the usable oxygen in blood, a person's hematocrit level is an important parameter when considering BOLD signal variations. Hematocrit levels could be unusually high in endurance athletes or folks who have just returned from an extended time living at altitude, for example. At the other extreme, we would be concerned about fMRI subjects experiencing anemia.

Levin et al. found a positive linear dependence of BOLD percent activation (BPA) on hematocrit level, the relationship being stronger in men than in women. Furthermore, "...9 men were studied before and following isotonic saline hemodilution, resulting in an average 6% reduction in hematocrit and an 8-31% reduction in BPA." This suggests checking whether your subject has just come off an i.v. saline drip because of food poisoning or flu.

Most causes of anemia are chronic and presumably you would be aware of them, but there are forms of anemia that might be of issue for BOLD studies in so-called normal volunteers. Recent utilization of i.v. fluids has just been mentioned while recent blood loss, perhaps from major surgery or giving blood, were considered in a prior section. Are these important? Knowing about them might permit you to interpret an outlier. Low hematocrit could also be relevant to studies of certain patient groups, e.g. acute traumatic brain injury, if polytrauma could have resulted in recent loss of blood.

Possible variations in hemoglobin suggests a role for normalization when comparing across groups (or even within subjects over an extended period of time). Lu, Yezhuvath & Xiao investigated the utility of baseline venous oxygenation as a normalizing parameter, finding that "...visual-evoked BOLD signal is significantly correlated with baseline venous T2 (P = 0.0003) and inclusion of physiologic modulator in the regression analysis can substantially reduce P values of group-level statistical tests."

I'll include here a special case. Sickle cell anemia (SCA) has profound effects on the BOLD response because resting CBF is increased. Remarkably, Bob Ogg's group was able to get fMRI responses in SCA children but the BOLD signals were diminished in amplitude. What is especially intriguing is that "... blood hemoglobin concentration and resting CBF were not predictive of BOLD signal amplitude in the SCA patients."No attempt was made to measure venous blood T2 in this particular study. Definitely an area ripe for more research.

Exercise

The long-term effects of exercise, particularly in endurance athletes, should be reflected in hematocrit level. What about the short-term effects of recent exercise on BOLD? If your subjects arrive at the scanner having just come from a boot camp class at the gym, or they've ridden a bike twenty miles from home, do you need to know about it?

A study by MacIntosh et al. showed that twenty minutes of aerobic exercise less than an hour before scanning decreased CBF in grey matter for up to 40 minutes. BOLD results from a go/no-go attention task were mostly consistent with the pre-exercise baseline. However, the same group used a similar exercise regimen prior to acquiring resting-state fMRI and "...observed a change in the resting-state BOLD functional connectivity of young healthy adults in three [resting state networks] RSNs, predominantly localized to cortical areas involved in sensorimotor activity." The latter study was limited in that "...at the time of the repeat rs-fMRI scan, the heart rate decreased after exercise but was nonetheless still significantly greater than the pre-exercise heart rate. Using heart rate as a covariate in the paired design group analysis did influence the session-related findings for two of the three significant RSNs. Blood pressure was not measured continuously so we cannot rule out the possibility that it too was elevated at the time of the repeat rs-fMRI scan."

There are clear opportunities for more research in this direction, but it would appear that we should insist on baseline measures of hematocrit (or venous T2) and blood pressure in addition to heart rate and respiration, or we risk misinterpreting BOLD changes.

(De)hydration

Using thermal exercise so that subjects dehydrated via sweating, versus a non-thermal exercise control, Kempton et al. found an increased fronto-parietal BOLD response during an executive function task while cognitive performance and CBF were unchanged. The authors suggest: "This pattern indicates that participants exerted a higher level of neuronal activity in order to achieve the same performance level." 

Caffeine

The effect of this vasoconstrictor on BOLD seem to have been studied more than any other drug (or foodstuff), probably because it's the most widely used stimulant in the world. There are many references to choose from, and what emerges is a picture best described as "it's complicated."

For a start it turns out that a subject's normal caffeine usage makes a difference if they're given a caffeine challenge. Laurienti et al. found "...that the BOLD signal change in visual cortex was significantly greater in high users than in low users in the presence of caffeine." Are your subjects high or low users? Would they even know? Here's a website listing many of the common dietary sources of caffeine. Many sources are well-known and obvious. Some, though, may surprise you.

Next there is the dose response. The amount of caffeine alters BOLD magnitude non-linearly. According to Chen & Parrish, the greatest effects on BOLD are associated with an intermediate caffeine dose of 2.5 mg/kg. (It must only be the neural response that goes all jittery with high doses!) In the early 2000s it was even suggested to use caffeine as a BOLD signal booster, but the non-linear dose response seems to have nixed that idea. Laurienti et al. puts it this way: "It is not possible to consistently enhance BOLD signal intensity magnitude by decreasing resting perfusion with caffeine."

In case the picture isn't sufficiently complicated already, it has been found that caffeine alters the temporal dynamics of visual BOLD responses. Intriguingly, administration of caffeine has also been found to enhance the linearity of the BOLD response to rapid visual stimuli. One assumes there must be a dose response that hasn't yet been investigated, however. What about resting-state fMRI and caffeine? For starters, caffeine changes resting-state connectivity for motor cortex. It also increases the variability of motor cortical connectivity.

So it appears that we may have an additional, potentially large, source of inter- (and intra?) subject variability based on the pattern of normal caffeine use, right? Curiously, the answer is ambiguous. Addicott, Peiffer & Laurienti investigated the effects of a caffeine dose (or placebo) on subjects representing a range of usage levels (including abstinence). Acute caffeine administration did produce measurable effects in BOLD, consistent with prior reports. But the changes weren't moderated either by normal use or by abstinence in regular users. The authors conclude "...that dietary caffeine use does not produce a significant effect on task-related BOLD activation."

Why might this be? Griffeth, Perthen & Buxton used a quantitative fMRI experiment to investigate caffeine's effects on BOLD and CBF simultaneously. They observed offsetting changes in baseline blood flow and oxygen metabolism with subsequent responses to visual stimuli, such that: "The combined effect was that BOLD responses pre- and post-caffeine were similar despite large underlying physiological changes, indicating that the magnitude of the BOLD response alone should not be interpreted as a direct measure of underlying neurophysiological changes."

Finally, let's assume you're interested in the neural consequences of caffeine use. How to differentiate from physiology? A good place to start is this 2010 review of caffeine's effects on cognition as well as BOLD. Then check out these more recent studies using simultaneous EEG-fMRI, comparing MEG to fMRI in the same subjects across separate sessions, or comparing fMRI to PET.

So, what's the bottom line regarding dietary caffeine as a potential BOLD confound? I don't know! But it sure looks like something that ought to be tracked, even if it's just self-reported accounts.

Consuming a stimulant in the land of depressants.

Alcohol

Ahhh, the world's favorite depressant. (The bulk of this post is being written overlooking vineyards in the Bennett Valley of Sonoma county, CA. Yes, that is coffee on the table above. Wine comes later, once the shadows point to bottom-right.) Let's deal with what we do know first. Acute administration of alcohol reduces BOLD activation in response to a visual stimulus, suppresses BOLD activity during a goal-directed visuomotor task, changes CBF in a dose-dependent manner, and modulates neurovascular coupling. Furthermore, acute alcohol consumption changes significantly several cognitive and visual networks mapped with a resting fMRI paradigm. (Also this study, on resting networks.) Finally, there are studies assessing the chronic effects of alcohol abuse, but I don't think chronic alcohol consumption can be considered as a potential physiologic confound because the neural effects will likely dominate.

Very surprisingly, nobody seems to have looked at the consequences of hangovers on fMRI. If acute alcohol administration causes changes to physiology and these changes are dose-dependent then at what point may we consider them negated? An hour? Eight hours? A night's sleep? Surely - surely? - someone must have looked at this. If not, it's a race to see who can get the first study reported.

Nicotine

The literature on nicotine, and cigarette smoking in particular, is a mixed bag as far as potential confounds go. In a study on nicotine-dependent smokers, Jacobsen et al. found no effect of intravenous nicotine on BOLD signals produced by visual stimulation. However, citing the "Considerable variability across individuals... in both the behavioral and fMRI blood oxygen level-dependent (BOLD) response to nicotine" that had been found in prior studies, Warbrick et al. observed: "...some participants showed an increase in activation while others showed a decrease in BOLD activation from the placebo to the [nasal] nicotine condition." Nineteen of the 41 total subjects were smokers.

So much for the controlled administration of nicotine in an experiment. What about the effects of smoking itself? Smoking may be a much larger concern because cigarettes produce carbon monoxide (CO), and hemoglobin has a high affinity for CO. If hemoglobin is carrying CO then it can't be carrying O2. How much of a concern is this displacement for BOLD? The literature doesn't say. Inhaling cigarette smoke is associated with a high number of health issues, including reduced pulmonary function, but I didn't find any literature addressing these concerns for BOLD-based fMRI.

One study from 2008 looked at heavy smoking as a potential confound in BOLD studies of schizophrenia, but found no difference with non-smokers using a sensorimotor task. However, another study from 2008 found significant differences in breath hold and visual activation tasks for heavy smokers versus controls. Why the contradiction? No idea I'm afraid.

Given the well documented respiratory and cardiac issues associated with chronic cigarette smoking, it seems to me that there is a very strong likelihood of systematic bias in the physiology of any group of smokers versus non-smokers. Would we compare marathon runners to couch potatoes and not expect significant differences in physiology? Surely, then, we should insist on some baseline measures of physiology, including perhaps pulmonary function tests (e.g. spirometry), or we should expect high inter-subject variability as well as smoker versus non-smoker group differences.

The physiologic effects produced by other nicotine delivery methods such as chewing tobacco, e-cigarettes (vaporizers) or patches may differ from the effects produced by cigarette smoking. Certainly that is my intuition but I wasn't able to locate any literature dealing with these issues. I think it suffices to say that we would want to uses caution when interpreting BOLD signals from smokers, and to predict systematic differences with non-smokers. That would just leave the dose effect (number of cigarettes/day, and perhaps depth and duration of smoke inhalation) to add a few more complications to the picture!

Illicit drugs

With the possible exception of cannabis (marijuana), we can probably discount most illegal (or partially legal) drugs as serious confounds for routine fMRI studies because they aren't in widespread use in the general population. (See Note 1.) But just in case, I did searches for several drugs, including ecstasy (MDMA), LSD and psychedelics in general, heroin, and methamphetamine, looking for direct investigations into possible confounding physiologic changes. Pattinson et al. studied the mu-opioid agonist remifentanil, finding only some small, regional modulations of the BOLD response to a hypercapnic challenge. Gollub et al.found that cocaine decreased CBF but BOLD responses to visual stimuli were normal. This doesn't imply that the underlying physiology is normal, however. As with caffeine, simultaneous changes in baseline CBF and oxygen metabolism could yield normal-looking responses in spite of different physiologic mechanisms. 

Cannabis is now in widespread use in most western populations. There are numerous studies of the (neural and behavioral) effects of this compound, but only a small handful of investigations of the physiologic effects relevant to BOLD.

There are a couple of literature reviews that have looked at neuroimaging and cannabis use. In a 2010 review the authors noted the general finding "...that resting global and prefrontal bloodflow are lower in cannabis users than in controls." An updated review, in 2013, assessed the findings as suggesting "...different patterns of resting global and brain activity during the performance of several cognitive tasks both in adolescents and adults, which may indicate compensatory effects in response to chronic cannabis exposure." Only neuroimaging studies involving chronic cannabis users were considered. Both reviews highlighted the methodological limitations of the work conducted to date and the considerable heterogeneity of results.

What sort of physiologic confounds might be of concern in cannabis users, anyway? Let's look at the acute effects of smoking marijuana. O'Leary et al. used O-15 PET and found: "Smoking marijuana resulted in intoxication, as assessed by a behavioral rating scale, but did not significantly alter mean behavioral performance on the attention task. Heart rate and blood pressure increased dramatically following smoking of marijuana but not placebo cigarettes. However, mean global CBF did not change significantly. Increased rCBF was observed in orbital and mesial frontal lobes, insula, temporal poles, anterior cingulate, as well as in the cerebellum. The increases in rCBF in anterior brain regions were predominantly in "paralimbic" regions and may be related to marijuana's mood-related effects. Reduced rCBF was observed in temporal lobe auditory regions, in visual cortex, and in brain regions that may be part of an attentional network (parietal lobe, frontal lobe and thalamus)."

That there are immediate effects on physiology following cannabis administration isn't all that surprising. And, as with acute alcohol administration, subjects smoking a joint immediately before doing an fMRI experiment probably aren't the prime concern. (If they are, you might want to check the effects of carbon monoxide, as for regular cigarettes.) What about the hangover effects or the chronic effects on pulmonary function?

In a 2006 study, Sneider et al. obtained dynamic susceptibility contrast (DSC) MRI (i.e. gadolinium bolus) from twelve "current, long-term daily cannabis users between 6 and 36 hr after the subjects' last reported cannabis use. Cannabis users demonstrated significantly increased blood volumes in the right frontal area (p < .05), in the left temporal area (p < .005), and in the cerebellum (p < .005) relative to comparison subjects."They followed that study with a longer duration of abstinence, scanning subjects with DSC MRI at 7 and 28 days after last cannabis use: "The present findings demonstrate that at Day 7, cannabis users continued to display increased blood volumes in the right frontal region, the left and right temporal regions, and the cerebellum. However, after 28 days of abstinence, only the left temporal area and cerebellum showed significantly increased CBV values in cannabis users. These findings suggest that while CBV levels begin to normalize with continued abstinence from cannabis, specifically in frontal areas, other temporal and cerebellar brain regions show slower CBV decreases."

Acute differences in regional CBF were also found in adolescent marijuana users, but the differences had resolved after four weeks of monitored abstinence. CBF wasn't measured at intermediate times, however, so we can't tell from this study how long any physiologic hangover effects might last. It also isn't clear whether the persistent effects are neural, physiologic or both, but you'd think that anything lasting more than a few days would have a strong neural basis. (This 2010 study observed BOLD signal changes in a spatial working memory task dependent on recency of use in adolescents.)

But do changes in baseline CBV and CBF imply different neurovascular coupling or altered BOLD responses? These might be more problematic for interpretation of fMRI data than persistent (neural) effects. According to Murphy et al., perhaps not. They used a finger-tapping task and determined no differences between groups of users of cocaine, nicotine or cannabis and control subjects.

Medications

Many legal (prescription or over-the-counter) analgesic and anti-inflammatory drugs don't seem to have been studied for their effects on neurovascular coupling. I couldn't find anything on aspirin (acetylsalicylic acid), paracetamol (acetaminophen), ibuprofen, codeine or oxycodone (oxycontin).  I also couldn't find any references investigating the undesirable effects of sedatives (except alcohol), Viagra (or its competitors), or antihistamines on BOLD physiology.

Of those medications that have been considered for possible physiologic confounds, the non-steroidal anti-inflammatory indomethacin was studied a decade ago and found to change BOLD and CBF but not CMRO2 for subjects conducting a simple motor task.

Blood pressure medications such as angiotensin, or anticoagulants such as warfarin, are presumably not something your subjects will use without telling you; your screening should catch these. Likewise asthma medications. I couldn't find any studies looking at potential physiologic confounds for these medications in any case. But Grichisch et al.investigated the potential effects of nasal insulin on CBF and BOLD and found no difference to a drug-free baseline. (Insulin and blood glucose level is considered again below, under Food.)

And then there are the so-called "smart drugs" and other brain enhancers such as methylphenidate (Ritalin), amphetamine (Adderall) and their ilk. (If I'm not using the correct collective terms for these compounds, I apologize.) Rao et al. assessed BOLD and perfusion changes for subjects performing a finger tapping task before and after oral methylphenidate and found no changes in neurovascular coupling. Otherwise, I wasn't able to find much literature addressing the potential for physiologic confounds. Marquand et al. looked at CBF changes due to methylphenidate while Nordin et al.assessed CBF changes due to amphetamine, both groups finding regional differences compared to placebo. But any implications for BOLD fMRI are unclear.

Literature coming out of the so-called pharmacological MRI (phMRI) field may provide the best clues for potential confounds to routine fMRI studies, should you have specific reason to be concerned about the use of medicines in your population of interest. Wang et al. reviewed the potential and challenges of using arterial spin labeling (ASL) in phMRI, "...with an emphasis on the methodologies used to control for potentially confounding vascular and systemic effects."

Anesthetics

It's unlikely you'll be surprised by use of anesthetics among your subjects. But there is certainly interest in scanning people who may have had recent surgery, e.g. traumatic brain-injured patients who may have required surgery to treat other injuries or even the TBI itself. You would presumably want to discount any residual neural and physiologic effects of anesthesia. For example, Qui et al. found "Low-dose sevoflurane significantly altered the task-induced CBF-BOLD coupling."Post surgery one expects altered blood pressure, heart rate and breathing rate, so there are plenty of options to measure and control for concomitant physiologic effects.

Foods

It seems that the biggest potential physiological confounds lie with those dietary substances known to alter neural activity, as already reviewed. In this section I'll try to deal with latent effects from things your subjects ingest for other reasons. For sustenance, say. Before that, though, what about the possible effects of restricted nutrition?

There are a couple of fMRI studies on fasting, one that found an effect in a motor task and another that found altered functional connectivity of visual cortex. There are also several reports of altered BOLD signal in response to experimental hypoglycemia (induced via insulin administration), including cognitive, visual stimulation, and sentence comprehension fMRI tasks. All three studies observed regional differences when compared to normal glucose levels. (See Note 2.) Patients with type I diabetes also showed different BOLD activity compared to controls. Additionally, there is one study using CBF and measures of oxidative metabolism (CMRO2) to assess the effects of hypoglycemia. They found that "...metabolism and flow remained coupled. Elementary motor task activation was not associated with any consistent larger activated flows. Thus it remains that although mild hypoglycemia induced an increase in basal flow and metabolism, a similar increase was not seen in task activation."Finally, glucose and fructose infusions were compared to saline and found to alter BOLD signal magnitude in opposite directions.

These are interesting studies, but you're probably more interested in any possible effect of blood glucose level on normal BOLD variability when your subjects aren't fasting intentionally or having their blood sugar levels manipulated artificially. I couldn't find any literature addressing this specific concern, however there is a very recent paper that compared an overnight fast with normal and hyperglycemic conditions. While the paper's abstract makes it sound like there were big changes in BOLD, in the Discussion they write: "These effects are comparatively small, yet may interfere with design sensitivity, when fasting status or blood glucose is not controlled in fMRI experiments."I only skimmed the paper but was left with the impression that the jury is still out on this one. Hyperglycemia, specifically, doesn't seem to be a concern for BOLD. Gruetter et al. observed no significant effects on BOLD for blood glucose of up to 300% of control levels.

Now let's move to what your subjects might be eating before getting a scan. In 2003 it was reported that ingesting lipid - 50 ml of canola oil - decreased BOLD response in a finger tapping task. But the potential effect of triglyceride levels in blood was measured again in 2009 by another group, whereupon no significant effects on BOLD were observed. Why the contradiction? We can only speculate. The authors of the 2009 study offer several plausible explanations in their discussion, and I am inclined to go with the more recent result pending further experimental data.

Then there is the antithesis to the high fat diet: salad. Aamand et al. used a nitrate challenge, "corresponding to the nitrate content of a large plate of salad," to trigger changes in BOLD reactivity without altering the baseline CBF. I really can't say whether this means we should be monitoring, even controlling, what our subjects eat before getting scanned.

Diurnal factors and sleep

Those of you studying anxiety or the effects of sleep or sleep deprivation are presumably acutely aware of and controlling for potentially confounding physiologic changes in your study designs. The effects of intentional manipulation of cortisol and sleep deficiency on fMRI are well documented. What about the effects in "normal volunteers?"Stress (level of blood cortisol) and amount of sleep (drowsiness) may be factors for the time of day your study is conducted, and they may vary with the populations you sample from. (Ditto for caffeine use!) But I can't find any references specifically addressing these possibilities. There is a now classic study which showed how inter-subject variability is much larger than intra-subject variability, so perhaps it's all baked in. Still, you would want to avoid biasing groups or longitudinal studies by the time of day you scan.

Hormones

I was unable to find any papers dealing with concomitant physiological changes accompanying the menstrual cycle. It seems reasonable to expect that significant effects might be reflected in the heart rate, respiration rate/depth and blood pressure. Direct vasodilatory or vasoconstrictive effects seem unlikely to me, but I'm no endocrinologist. There are papers dealing with performance differences across menstrual cycle, but if these effects repeat they may have a neural basis rather than being physiological artifacts.

Women (and soon men?) going on or off hormonal contraceptives during a study might be a bigger deal. A recent study found a marked change in BOLD responses for women using oral contraception. The effects could be neural or concomitant physiological changes, or a mixture of both.

I couldn't leave this section without mentioning oxytocin. It's quite in vogue. I wasn't able to find any studies that had have investigated systemic physiological effects of this compound as they pertain to BOLD. But giving subjects oxytocin - or any other pharmacological agent come to that - without checking for concomitant physiological changes is taking a leap of faith that all effects are neural. At bare minimum I would expect blood pressure to be recorded with and without the challenge, and then I would want to see heart rate and respiration recorded during fMRI.

Age and disease

These variables should come as no major surprise if you are screening your subjects in any way at all. But in order to be complete and to give you a place to start if you're doing group studies across the lifespan, or comparing disease states to normal volunteers, I offer these papers:

Alterations in the BOLD fMRI signal with aging and disease: a challenge for neuroimaging.

Age-related differences in memory-encoding fMRI responses after accounting for decline in vascular reactivity.

Evidence that neurovascular coupling underlying the BOLD effect increases with age during childhood.

Age-related differences in cerebral blood flow underlie the BOLD fMRI signal in childhood.

Age dependence of hemodynamic response characteristics in human functional magnetic resonance imaging. 

Neural mechanisms of age-related slowing: the ΔCBF/ΔCMRO2 ratio mediates age-differences in BOLD signal and human performance.

Calibrated fMRI during a cognitive Stroop task reveals reduced metabolic response with increasing age.

Visualization of altered neurovascular coupling in chronic stroke patients using multimodal functional MRI.

Acceleration

If your subjects have just returned from a six-month stint on the International Space Station and you want to compare fMRI before and after the flight, you are no doubt fully aware of the many physiological factors to take into account where extended microgravity is concerned. Short duration microgravity - say, the few minutes to be experienced by lucky punters taking Virgin Galactic flights in the not-too-distant future - is probably not much of a concern, unless the flight happens the morning of their fMRI scan.

At the other end of the scale are those people exposing themselves to high acceleration; much greater than 1 g. Presumably these exposures are seconds or at most minutes in duration, but if the exposure occurs frequently or happens in the hours before participating in an fMRI experiment, we should be sure to exclude delayed physiologic changes. To date, though, I can find no reports detailing possible confounding effects of aerobatic plane flights, gymnastics, fairground rides, trampolining and the like. I am assuming these aren't going to be common issues for the vast majority of fMRI studies, but concomitant physiologic changes would be worth consideration if your research happens to involve fighter pilots, race car drivers, etc. Presumably measuring blood pressure, respiration and heart rates would capture the main effects. 





Notes:

1.  The other day someone told me about a subject who had answered negative to a drug screen but who then got very concerned when, on the morning of the scan, she was asked to give a urine sample. (I don't recall why the urine sample was required, and presumably it was stated on the recruitment flier which I can only assume the subject didn't read fully before applying to do the experiment. In any event, it was a vital part of the experiment and not a law-enforcement action!) She apparently returned from the bathroom with a sample of tap water in her specimen tube. The point? People will lie on screening forms if they are trying to make the $20 an hour you're offering. So if your experiment is highly sensitive to illicit drug use then you may want to include the option (or the threat?) of a urine test to help filter those people out.

2.  Hypoglycemia was also one of the conditions used (in rats) by Ogawa et al.in their preliminary demonstration of the BOLD effect. From the abstract of their 1990 paper: "This blood oxygenation level-dependent (BOLD) contrast follows blood oxygen changes induced by anesthetics, by insulin-induced hypoglycemia, and by inhaled gas mixtures that alter metabolic demand or blood flow."So we're not exactly covering new ground here, are we?

There was, of course, a huge procedural failure that allowed a large, magnetic oxygen cylinder into the MRI facility in the first place. No doubt the investigation will find ample blame to spread around. But the solution to the problem is rather simple: education/training coupled with standard operating procedures to nix the threat. As procedures go it's not especially difficult. (By comparison, over 34,000 people manage to get themselves killed on US roads every single year. Clearly, we can't drive for shit. Our procedures are severely wanting in this department.) And if you're ever in doubt as to whether an item can be brought safely into the MRI suite there is always - always! - someone you can go to for an expert opinion. In my facility no equipment is allowed through the door without that expert opinion being cast.

So let's shift to the part of this fiasco that really got my attention: the claim that the magnet quench circuit malfunctioned. From the second article, above:

"At a press conference on Wednesday, a day after this newspaper broke the story, senior officials of Tata Memorial-run Advance Centre or Treatment Research and Education in Cancer (ACTREC) in Khargar said that because a switch to disable the machine's magnetic field malfunctioned, it took engineers four hours to disengage the two employees - a ward boy and a technician -- stuck to the machine, when it should not have taken more than 30 seconds."

Then, later in the same article:

"Dr Gupta said the MRI machine was bought just four years back and its last periodic maintenance was carried out by GE engineers just six months back. "We were shocked. Despite several attempts to switch off the machine, we just could not disable it. Ideally, we should have removed both of our employees in 30 seconds, but they were stuck for four hours," he said.

Dr Gupta said that while ACTREC engineers had reached the spot within minutes of the mishap, they just could not switch off the machine. The manufacturer was informed immediately, but the GE engineer who arrived after an hour too could not turn the machine off. Nearly two-and-a-half hours after the incident, three more GE engineers arrived and replaced the existing electronic circuit with a new one and then switched off the machine."

So I asked my local service engineer - from Siemens, but I would wager large sums of money that I'd have got a similar response from any other vendor - what guarantees I have that my magnet will quench in the event I need to push the Big Red Button. Here is an edited version of his unequivocal response:

"Your rundown unit will work.  It is tested thoroughly every year, parts are tested daily, and parts are tested on the fly (always).

A magnet load simulator tool is connected (in place of the magnet) every year, and the quench circuits are tested into these loads to verify function.  Yes, we push the quench buttons every year. 

In case of power failure, there is a battery backup for the rundown unit.  The battery is load tested automatically every day at 2 AM to insure it is sustaining a proper charge and capacity (if it fails this test, the scanner will NOT SCAN!).

There is a small amount of current that is always applied through the quench and switch heaters in the magnet.  Seems like a scary concept, but it is shown not to produce enough heat to activate the heaters.  With this current applied, the voltage drop across the heaters is proportional to their resistance; if this value exceeds the allowed specification (too low or too high), you get an alarm.  You heard this alarm not too long ago when the switch/quench heater connecter was corroded/wet from condensation, and this resistance value was incorrect as a consequence.  So you got first-hand experience of what happens if the heaters aren’t connected with the correct resistance.  Further, this is also tested every year . . . we simulate high + low resistances with a tool for all three heaters in the magnet to check the monitoring circuits for proper function.

We even disconnect one of the ‘quench buttons’ to be sure the system gives an alarm reporting ‘too few buttons’.  Yes, even the presence of the switch being electrically connected is monitored.

These tests are monitored with a date/time stamp (file C:\MedCom\log\MsupHistory.log).  This data is transferred via remote connection to our server, and monitored by Erlangen (automated running script, I believe).  When the Siemens Magnet experts (not Oxford!) visit for magnet service, they also download this data from the magnet monitoring system.  If the logs show this test hasn’t been run within an approx 12 month interval, a service call is opened and an engineer is dispatched to check this immediately.

Again, your rundown unit will work.  The heaters are monitored and known to be connected with the correct resistance, the switches are monitored as being physically connected, the rundown unit ‘quenches’ a simulated magnet, and the battery backup is tested, and the monitoring circuits are all tested.  And we have a quality process in place to be sure these tests/functions are carried out."

I'm happy enough with this answer, but as someone who always likes backups for everything I still asked whether there was another way to quench the magnet, just in case. I was informed that there are other ways (and I already knew of one of them) but they are rather scary and not recommended for a variety of valid safety reasons. (Personnel safety, not magnet safety.)  They might be used in drastic cases, like perhaps if a fire destroys all the quench circuitry. But this is a very different scenario than trying to rescue someone stuck to the magnet. For everyday safety I am reassured that my quench circuit will activate if the button is depressed.

Whether or not GE was really at fault in Mumbai we shall learn eventually, I hope. (I have heard rumors that some sites like to bypass their quench circuit in order to avoid having the cost of recharging the magnet should the quench button get activated. Insert your own exclamations of disbelief here because I'm incredulous.) In the mean time, this sorry saga is an opportunity for all of us to review our own procedures and take the extra moments to ensure that we've done everything humanely possible to eliminate risks. There really is no excuse.

• Added DRIFTER to the list of de-noising methods
• Added a reference for sex differences in hematocrit and the effects on BOLD
• Added several medication classes, including statins, sedatives & anti-depressants
• Added a few dietary supplements, under Food

A recent conversation on Twitter led to the suggestion that someone compile a list of physiological effects of concern for BOLD. That is, a list of potentially confounding physiological changes that could arise sympathetically in an fMRI experiment, such as altered heart rate due to the stress of a task, or that could exist as a systematic difference between groups. What follows is the result of a PubMed literature search (mostly just the abstracts) where I have tried to identify either recent review articles or original research that can be used as starting points for learning more about candidate effects. Hopefully you can then determine whether a particular factor might be of concern for your experiment.

This is definitely not a comprehensive list of all literature pertaining to all potential physiological confounds in fMRI, and I apologize if your very important contribution didn't make it into the post. Also, please note that I am not a physiologist so if I go seriously off piste in interpreting the literature, please forgive me and then correct my course. I would like to hear from you (comments below, or via Twitter) if I have omitted critical references or effects from the list, or if I have misinterpreted something. As far as possible I've tried to restrict the review to work in humans unless there was nothing appropriate, in which case I've included some animal studies if I think they are directly relevant. I'll try to keep this post up-to-date as new studies come out and as people let me know about papers I've missed.

A final caution before we begin. It occurs to me that some people will take this list as (further) proof that all fMRI experiments are hopelessly flawed and will use it as ammunition. At the other extreme there will be people who see this list as baseless scare-mongering. How you use the list is entirely up to you, but my intent is to provide cautious fMRI scientists with a mechanism to (re)consider potential physiologic confounds in their experiments, and perhaps stimulate the collection of parallel data that might add power to those experiments.

Getting into BOLD physiology

There are some good recent articles that introduce the physiological artifacts of prime concern. Tom Liu has reviewed neurovascular factors in resting-state functional MRI and shows how detectable BOLD signals arise from physiological changes in the first place. Kevin Murphy et al. then review some of the most common confounds in resting-state fMRI and cover a few ways these spurious signal changes can be characterized and even removed from data. Finally, Dan Handwerker et al. consider some of the factors causing hemodynamic variations within and, in particular, between subjects.

Once you start really looking into this stuff it can be hard not to get despondent. Think of the large number of potential manipulations as opportunities, not obstacles! Perhaps let The Magnetic Fields get you in the mood with their song, "I don't like your (vascular) tone." Then read on. It's a long list.

Breathing and heart rates

These are undoubtedly the two big issues and will be top-of-mind for most people. Murphy et al. do a good job introducing the topic of cleaning raw data, e.g. using independent respiration and heart rate data. There are numerous correction methods available today (see the section, "Time series cleanup with physiological recordings" in the Murphy paper) and these references are perhaps the best places to start if you want to dig into the origins of the problems. In reverse chronological order we have:

PHYCAA+ by Churchill & Strother
GLMdenoise by Kay, Rokem, Winawer, Dougherty & Wandell.
RIPTiDe by Frederick, Nickerson & Tong (using near-infrared spectroscopy, NIRS)
DRIFTER by Särkkä, Solin, Nummenmaa, Vehtari, Auranen, Vanni & Lin.
RVT-HR by Chang, Cunningham & Glover
RVT by Birn, Diamond, Smith & Bandettini
RETROICOR by Glover, Li & Reese

Then there are also numerous ICA-based methods, including multi-echo acquisition with a model fit. I won't comment on the relative or absolute performance of any of these schemes, just let you know you have options! But one important lesson seems to be this: measuring heart rate and respiration independently isn't likely to hurt.

I'll also mention the inter-relatedness of HR and breathing, through vagal tone. This has implications for fMRI studies employing emotional stimuli.

CO2: hypercapnia and hypocapnia

Inspired CO2 is a well-known and powerful vasodilator. Similarly, breath holding is known to lead to hypercapnia (increases in arterial CO2) and thus vasodilation. Thus, under normal conditions (no external CO2 source) the arterial concentration of CO2 will be closely linked to the breathing rate, as demonstrated in this 2004 study. The quantitative relationship between breathing rate and end-tidal CO2 is further investigated in this 2009 study.

As for dealing with the variation, in a 2012 study it was demonstrated that clamping arterial CO2 to a narrow range improved retrospective physiological artifact removal from resting-state and task-based data. I like this idea because I've often wondered whether not having the magnet bore properly ventilated could cause hypercapnia. I can't find any papers that have looked at this issue, however. The closest I've found is a study that used CO2 supplied artificially, clearly demonstrating the potential interaction of inspired CO2 with stimulus-driven BOLD. (There was an earlier study in rats that reached essentially the same conclusion.) It's probably best to keep the ventilation constant in the magnet bore for all subjects, just in case.

It turns out there may be gender bias, too. A 2010 study found that men have higher cerebrovascular reactivity to a CO2 challenge. So, not only do men tend to exhibit more head movement than women on average, men also appear to be more sensitive to CO2.

O2: hyperoxia and hypoxia

While refluxed CO2 may be a legitimate concern, incidental hyperoxia seems unlikely unless your MRI facility has a very strange design indeed. But, if you happen to be doing fMRI of divers, high-altitude pilots or astronauts in a simulated environment, do bear in mind the possibility of arterial BOLD from dissolved O2 if your subjects are breathing 100% oxygen. I suppose this ought to be a concern in anesthetized fMRI subjects, too. (I'll consider the physiologic consequences of anesthetics below.) If you're comparing normal volunteers (awake) to anesthetized patients then you might want to consider giving the former group the same gas mixture (sans anesthetic) as the latter.

What about hypoxia? If you have a well-designed MRI facility then ordinarily acute hypoxia shouldn't be an issue. But if your scanner is in a poorly ventilated basement you might want to consider an oxygen sensor, assuming you don't already have one, or improving your air conditioning. Otherwise you may be contributing to the variability in your subjects' BOLD responses.

Chronic hypoxia may be an important issue to track, too. Yan et al. studied immigrants who had grown up at high altitude (HA, 2616–4200 m) and relocated to sea level (SL, < 400 m) to a group of sea-level controls, and found both anatomical and resting state fMRI differences associated with "chronic high altitude hypoxia." These changes were in spite of "...no significant differences in hemoglobin levels, circulating red blood cell count, blood pressure, and pulse rate between HA residents and SL controls." In a separate study, Yan et al. found persistent, correlated alterations in verbal working memory and BOLD responses for subjects who had relocated from HA (2,982.8 ± 478.7 m) to SL (< 400 m), even though they had been resident at sea level for at least one year and in spite of consistent hemoglobin levels.

Vasomotion

According to that great oracle, Wikipedia, "vasomotion is the spontaneous oscillation in tone of blood vessels, independent of heart beat, innervation or respiration." Fair enough. Our concern is whether vasomotion might differ across groups, or perhaps might even vary for an individual under different conditions. Again according to Wikipedia, "vasomotion has been shown to be altered in a variety of pathological situations, with vessels from both hypertensive and diabetic patients displaying altered flow patterns as compared to normotensive vessels."

Murphy et al. say this: "If vasomotion is independent of cardiac, respiratory, arterial CO2 concentration and blood pressure fluctuations, its low-frequency characteristics will present another confound for resting-state fMRI BOLD oscillations."

How serious a concern might spontaneous low-frequency variations be for your average BOLD study? A recent paper by Tong & Frederick claims that low frequency oscillations (LFOs) are a larger component of physiologic noise than respiration or heart rate, even when respiration and heart rate frequencies are aliased because of the relatively long TRs that are typical of most fMRI studies. There is also evidence from intraoperative recordings that vasomotion (at ~0.1 Hz) is "spatially localized to distinct regions of the cortex, exhibited wave-like propagation, and involved oscillations in the diameter of specific pial arterioles, indicating that the effect was not the result of systemic blood pressure oscillations."Not small potatoes, then.

What do these LFOs tell us, and (how) do they relate to the ongoing intrinsic neural activity that we assume is driving resting-state fMRI? Methinks this might become the next fashionable study area, if it's not already.

Blood pressure

A good introduction to blood pressure (BP) and its relationship to CBF is given in the Murphy, Birn & Bandettini review. We might be interested in the subject's baseline (pre-scan) BP, but we might also be interested in the BP during a time series acquisition given that BP changes when laying supine versus standing, during stress, etc.

I've been unable to locate any papers showing whether TR-to-TR (real-time) BP is of any use in explaining variance in BOLD time series data. Baseline BP (before MRI) was found to offer only a small normalizing effect on visual-evoked BOLD signals when tested across two conditions; baseline venous T2 was considerably better. A study by Lui et al. in cocaine-dependent (human) subjects found that dobutamine infusion raised mean arterial BP (MAP) but produced only localized BOLD signal changes (in anterior cingulate) that correlated with the BP rise. Gianaros et al. observed a correlation between mean arterial BP and BOLD activity in several brain regions of participants conducting a stressful Stroop task, with BP measured once for each block of sixteen 90-second task blocks.

This is Murphy et al.'s take:

"Evidence of the influence of blood pressure oscillations on resting-state fMRI fluctuations in humans is sparse. Blood pressure levels in rats have been shown to affect evoked fMRI responses, with transient hypertension increasing BOLD (Wang et al., 2006) and CBF (Qiao et al., 2007) signals. Under hypotension, neural activity-evoked CBV increases in visual cortex are negligible compared to ~ 10% at normal blood pressure levels (Nagaoka et al., 2006). Increases in the amplitude of low-frequency BOLD fluctuations have been demonstrated with a drop in mean arterial pressure (Biswal and Kannurpatti, 2009). As supporting evidence in humans, BOLD signal correlates of heart rate and pulse height in the low frequency range have been discovered with fluctuations in cardiac rate explaining up to 11% of the variance in the resting-state BOLD signal (Chang et al., 2009, de Munck et al., 2008 and Shmueli et al., 2007)."

What about chronic hypertension? Another study from Gianaros et al. used the Stroop test to stress healthy young volunteers as a model for determining whether reactivity to psychological stressors might be used to assess risk from hypertension. They found: "Individuals exhibiting greater stressor-evoked MAP reactivity showed (1) greater amygdala activation, (2) lower amygdala gray matter volume, and (3) stronger positive functional connectivity between the amygdala and perigenual anterior cingulate cortex and brainstem pons."

Hypotension may be a factor in cases of major blood loss, perhaps including recent blood donation. While I wasn't able to find any papers looking at recent blood loss (or donation) on BOLD fMRI, Kalisch et al. used a hemorrhage model in rats and observed heterogeneous correlations between BOLD and BP. They suggest that "...a BOLD decrease during a decrease in BP may result from an increase in mostly venous cerebral blood volume (CBV) as an autoregulatory response to maintain cerebral blood flow (CBF) during decreased perfusion pressure."

Hematocrit level

Hematocrit level is the percentage of red blood cells in the blood. And since red blood cells carry hemoglobin, which itself transports the usable oxygen in blood, a person's hematocrit level is an important parameter when considering BOLD signal variations. Hematocrit levels could be unusually high in endurance athletes or folks who have just returned from an extended time living at altitude, for example. At the other extreme, we would be concerned about fMRI subjects experiencing anemia.

Levin et al. found a positive linear dependence of BOLD percent activation (BPA) on hematocrit level, the relationship being stronger in men than in women. Furthermore, "...9 men were studied before and following isotonic saline hemodilution, resulting in an average 6% reduction in hematocrit and an 8-31% reduction in BPA." This suggests checking whether your subject has just come off an i.v. saline drip because of food poisoning or flu.

There are also well-known sex differences for hematocrit level, with men having an average value several percent (5% seems to be the consensus figure) higher than women. A poster by Yang et al. at this year's HBM conference showed how variation in hematocrit across individuals produced a systematic difference in several measures derived from resting-state fMRI.

Most causes of anemia are chronic and presumably you would be aware of them, but there are forms of anemia that might be of issue for BOLD studies in so-called normal volunteers. Recent utilization of i.v. fluids has just been mentioned while recent blood loss, perhaps from major surgery or giving blood, were considered in a prior section. Are these important? Knowing about them might permit you to interpret an outlier. Low hematocrit could also be relevant to studies of certain patient groups, e.g. acute traumatic brain injury, if polytrauma could have resulted in recent loss of blood.

Possible variations in hemoglobin suggests a role for normalization when comparing across groups (or even within subjects over an extended period of time). Lu, Yezhuvath & Xiao investigated the utility of baseline venous oxygenation as a normalizing parameter, finding that "...visual-evoked BOLD signal is significantly correlated with baseline venous T2 (P = 0.0003) and inclusion of physiologic modulator in the regression analysis can substantially reduce P values of group-level statistical tests."

I'll include here a special case. Sickle cell anemia (SCA) has profound effects on the BOLD response because resting CBF is increased. Remarkably, Bob Ogg's group was able to get fMRI responses in SCA children but the BOLD signals were diminished in amplitude. What is especially intriguing is that "... blood hemoglobin concentration and resting CBF were not predictive of BOLD signal amplitude in the SCA patients."No attempt was made to measure venous blood T2 in this particular study. Definitely an area ripe for more research.

Exercise

The long-term effects of exercise, particularly in endurance athletes, should be reflected in hematocrit level. What about the short-term effects of recent exercise on BOLD? If your subjects arrive at the scanner having just come from a boot camp class at the gym, or they've ridden a bike twenty miles from home, do you need to know about it?

A study by MacIntosh et al. showed that twenty minutes of aerobic exercise less than an hour before scanning decreased CBF in grey matter for up to 40 minutes. BOLD results from a go/no-go attention task were mostly consistent with the pre-exercise baseline. However, the same group used a similar exercise regimen prior to acquiring resting-state fMRI and "...observed a change in the resting-state BOLD functional connectivity of young healthy adults in three [resting state networks] RSNs, predominantly localized to cortical areas involved in sensorimotor activity." The latter study was limited in that "...at the time of the repeat rs-fMRI scan, the heart rate decreased after exercise but was nonetheless still significantly greater than the pre-exercise heart rate. Using heart rate as a covariate in the paired design group analysis did influence the session-related findings for two of the three significant RSNs. Blood pressure was not measured continuously so we cannot rule out the possibility that it too was elevated at the time of the repeat rs-fMRI scan."

There are clear opportunities for more research in this direction, but it would appear that we should insist on baseline measures of hematocrit (or venous T2) and blood pressure in addition to heart rate and respiration, or we risk misinterpreting BOLD changes.

(De)hydration

Using thermal exercise so that subjects dehydrated via sweating, versus a non-thermal exercise control, Kempton et al. found an increased fronto-parietal BOLD response during an executive function task while cognitive performance and CBF were unchanged. The authors suggest: "This pattern indicates that participants exerted a higher level of neuronal activity in order to achieve the same performance level." 

Caffeine

The effect of this vasoconstrictor on BOLD seem to have been studied more than any other drug (or foodstuff), probably because it's the most widely used stimulant in the world. There are many references to choose from, and what emerges is a picture best described as "it's complicated."

For a start it turns out that a subject's normal caffeine usage makes a difference if they're given a caffeine challenge. Laurienti et al. found "...that the BOLD signal change in visual cortex was significantly greater in high users than in low users in the presence of caffeine." Are your subjects high or low users? Would they even know? Here's a searchable website containing many of the common dietary sources of caffeine. Many sources are well-known and obvious, e.g. dark chocolate. Some, though, may surprise you.

Next there is the dose response. The amount of caffeine alters BOLD magnitude non-linearly. According to Chen & Parrish, the greatest effects on BOLD are associated with an intermediate caffeine dose of 2.5 mg/kg. (It must only be the neural response that goes all jittery with high doses!) In the early 2000s it was even suggested to use caffeine as a BOLD signal booster, but the non-linear dose response seems to have nixed that idea. Laurienti et al. puts it this way: "It is not possible to consistently enhance BOLD signal intensity magnitude by decreasing resting perfusion with caffeine."

In case the picture isn't sufficiently complicated already, it has been found that caffeine alters the temporal dynamics of visual BOLD responses. Intriguingly, administration of caffeine has also been found to enhance the linearity of the BOLD response to rapid visual stimuli. One assumes there must be a dose response that hasn't yet been investigated, however. What about resting-state fMRI and caffeine? For starters, caffeine changes resting-state connectivity for motor cortex. It also increases the variability of motor cortical connectivity.

So it appears that we may have an additional, potentially large, source of inter- (and intra?) subject variability based on the pattern of normal caffeine use, right? Curiously, the answer is ambiguous. Addicott, Peiffer & Laurienti investigated the effects of a caffeine dose (or placebo) on subjects representing a range of usage levels (including abstinence). Acute caffeine administration did produce measurable effects in BOLD, consistent with prior reports. But the changes weren't moderated either by normal use or by abstinence in regular users. The authors conclude "...that dietary caffeine use does not produce a significant effect on task-related BOLD activation."

Why might this be? Griffeth, Perthen & Buxton used a quantitative fMRI experiment to investigate caffeine's effects on BOLD and CBF simultaneously. They observed offsetting changes in baseline blood flow and oxygen metabolism with subsequent responses to visual stimuli, such that: "The combined effect was that BOLD responses pre- and post-caffeine were similar despite large underlying physiological changes, indicating that the magnitude of the BOLD response alone should not be interpreted as a direct measure of underlying neurophysiological changes."

Finally, let's assume you're interested in the neural consequences of caffeine use. How to differentiate from physiology? A good place to start is this 2010 review of caffeine's effects on cognition as well as BOLD. Then check out these more recent studies using simultaneous EEG-fMRI, comparing MEG to fMRI in the same subjects across separate sessions, or comparing fMRI to PET.

So, what's the bottom line regarding dietary caffeine as a potential BOLD confound? I don't know! But it sure looks like something that ought to be tracked, even if it's just self-reported accounts.

Consuming a stimulant in the land of depressants.

Alcohol

Ahhh, the world's favorite depressant. (The bulk of this post is being written overlooking vineyards in the Bennett Valley of Sonoma county, CA. Yes, that is coffee on the table above. Wine comes later, once the shadows point to bottom-right.) Let's deal with what we do know first. Acute administration of alcohol reduces BOLD activation in response to a visual stimulus, suppresses BOLD activity during a goal-directed visuomotor task, changes CBF in a dose-dependent manner, and modulates neurovascular coupling. Furthermore, acute alcohol consumption changes significantly several cognitive and visual networks mapped with a resting fMRI paradigm. (Also this study, on resting networks.) Finally, there are studies assessing the chronic effects of alcohol abuse, but I don't think chronic alcohol consumption can be considered as a potential physiologic confound because the neural effects will likely dominate.

Very surprisingly, nobody seems to have looked at the consequences of hangovers on fMRI. If acute alcohol administration causes changes to physiology and these changes are dose-dependent then at what point may we consider them negated? An hour? Eight hours? A night's sleep? Surely - surely? - someone must have looked at this. If not, it's a race to see who can get the first study reported.

Nicotine

The literature on nicotine, and cigarette smoking in particular, is a mixed bag as far as potential confounds go. In a study on nicotine-dependent smokers, Jacobsen et al. found no effect of intravenous nicotine on BOLD signals produced by visual stimulation. However, citing the "Considerable variability across individuals... in both the behavioral and fMRI blood oxygen level-dependent (BOLD) response to nicotine" that had been found in prior studies, Warbrick et al. observed: "...some participants showed an increase in activation while others showed a decrease in BOLD activation from the placebo to the [nasal] nicotine condition." Nineteen of the 41 total subjects were smokers.

So much for the controlled administration of nicotine in an experiment. What about the effects of smoking itself? Smoking may be a much larger concern because cigarettes produce carbon monoxide (CO), and hemoglobin has a high affinity for CO. If hemoglobin is carrying CO then it can't be carrying O2. How much of a concern is this displacement for BOLD? The literature doesn't say. Inhaling cigarette smoke is associated with a high number of health issues, including reduced pulmonary function, but I didn't find any literature addressing these concerns for BOLD-based fMRI.

One study from 2008 looked at heavy smoking as a potential confound in BOLD studies of schizophrenia, but found no difference with non-smokers using a sensorimotor task. However, another study from 2008 found significant differences in breath hold and visual activation tasks for heavy smokers versus controls. Why the contradiction? No idea I'm afraid.

Given the well documented respiratory and cardiac issues associated with chronic cigarette smoking, it seems to me that there is a very strong likelihood of systematic bias in the physiology of any group of smokers versus non-smokers. Would we compare marathon runners to couch potatoes and not expect significant differences in physiology? Surely, then, we should insist on some baseline measures of physiology, including perhaps pulmonary function tests (e.g. spirometry), or we should expect high inter-subject variability as well as smoker versus non-smoker group differences.

The physiologic effects produced by other nicotine delivery methods such as chewing tobacco, e-cigarettes (vaporizers) or patches may differ from the effects produced by cigarette smoking. Certainly that is my intuition but I wasn't able to locate any literature dealing with these issues. I think it suffices to say that we would want to uses caution when interpreting BOLD signals from smokers, and to predict systematic differences with non-smokers. That would just leave the dose effect (number of cigarettes/day, and perhaps depth and duration of smoke inhalation) to add a few more complications to the picture.

Illicit drugs

With the possible exception of cannabis (marijuana), we can probably discount most illegal (or partially legal) drugs as serious confounds for routine fMRI studies because they aren't in widespread use in the general population. (See Note 1.) But just in case, I did searches for several drugs, including ecstasy (MDMA), LSD and psychedelics in general, heroin, and methamphetamine, looking for direct investigations into possible confounding physiologic changes. Pattinson et al. studied the mu-opioid agonist remifentanil, finding only some small, regional modulations of the BOLD response to a hypercapnic challenge. Gollub et al.found that cocaine decreased CBF but BOLD responses to visual stimuli were normal. This doesn't imply that the underlying physiology is normal, however. As with caffeine, simultaneous changes in baseline CBF and oxygen metabolism could yield normal-looking responses in spite of different physiologic mechanisms. 

Cannabis is now in widespread use in most western populations. There are numerous studies of the (neural and behavioral) effects of this compound, but only a small handful of investigations of the physiologic effects relevant to BOLD.

There are a couple of literature reviews that have looked at neuroimaging and cannabis use. In a 2010 review the authors noted the general finding "...that resting global and prefrontal bloodflow are lower in cannabis users than in controls." An updated review, in 2013, assessed the findings as suggesting "...different patterns of resting global and brain activity during the performance of several cognitive tasks both in adolescents and adults, which may indicate compensatory effects in response to chronic cannabis exposure." Only neuroimaging studies involving chronic cannabis users were considered. Both reviews highlighted the methodological limitations of the work conducted to date and the considerable heterogeneity of results.

What sort of physiologic confounds might be of concern in cannabis users, anyway? Let's look at the acute effects of smoking marijuana. O'Leary et al. used O-15 PET and found: "Smoking marijuana resulted in intoxication, as assessed by a behavioral rating scale, but did not significantly alter mean behavioral performance on the attention task. Heart rate and blood pressure increased dramatically following smoking of marijuana but not placebo cigarettes. However, mean global CBF did not change significantly. Increased rCBF was observed in orbital and mesial frontal lobes, insula, temporal poles, anterior cingulate, as well as in the cerebellum. The increases in rCBF in anterior brain regions were predominantly in "paralimbic" regions and may be related to marijuana's mood-related effects. Reduced rCBF was observed in temporal lobe auditory regions, in visual cortex, and in brain regions that may be part of an attentional network (parietal lobe, frontal lobe and thalamus)."

That there are immediate effects on physiology following cannabis administration isn't all that surprising. And, as with acute alcohol administration, subjects smoking a joint immediately before doing an fMRI experiment probably aren't the prime concern. (If they are, you might want to check the effects of carbon monoxide, as for regular cigarettes.) What about the hangover effects or the chronic effects on pulmonary function?

In a 2006 study, Sneider et al. obtained dynamic susceptibility contrast (DSC) MRI (i.e. gadolinium bolus) from twelve "current, long-term daily cannabis users between 6 and 36 hr after the subjects' last reported cannabis use. Cannabis users demonstrated significantly increased blood volumes in the right frontal area (p < .05), in the left temporal area (p < .005), and in the cerebellum (p < .005) relative to comparison subjects."They followed that study with a longer duration of abstinence, scanning subjects with DSC MRI at 7 and 28 days after last cannabis use: "The present findings demonstrate that at Day 7, cannabis users continued to display increased blood volumes in the right frontal region, the left and right temporal regions, and the cerebellum. However, after 28 days of abstinence, only the left temporal area and cerebellum showed significantly increased CBV values in cannabis users. These findings suggest that while CBV levels begin to normalize with continued abstinence from cannabis, specifically in frontal areas, other temporal and cerebellar brain regions show slower CBV decreases."

Acute differences in regional CBF were also found in adolescent marijuana users, but the differences had resolved after four weeks of monitored abstinence. CBF wasn't measured at intermediate times, however, so we can't tell from this study how long any physiologic hangover effects might last. It also isn't clear whether the persistent effects are neural, physiologic or both, but you'd think that anything lasting more than a few days would have a strong neural basis. (This 2010 study observed BOLD signal changes in a spatial working memory task dependent on recency of use in adolescents.)

But do changes in baseline CBV and CBF imply different neurovascular coupling or altered BOLD responses? These might be more problematic for interpretation of fMRI data than persistent (neural) effects. According to Murphy et al., perhaps not. They used a finger-tapping task and determined no differences between groups of users of cocaine, nicotine or cannabis and control subjects.

Medications

Many legal (prescription or over-the-counter) analgesic and anti-inflammatory drugs don't seem to have been studied for their effects on neurovascular coupling. I couldn't find anything on aspirin (acetylsalicylic acid), paracetamol (acetaminophen), ibuprofen, codeine or oxycodone (oxycontin).  I also couldn't find any references investigating the undesirable effects of Viagra (or its competitors) or antihistamines on BOLD physiology. (One group has even suggested using diphenhydramine, active ingredient in the US/Canadian version of the antihistamine drug, Benadryl as a way to combat nausea and vertigo in 7 T MRI.)

Sedatives have been used in several fMRI studies, but to date there have been no investigations into concomitant physiologic effects for the popular benzodiazepines such as diazepam (Valium), lorazepam (Ativan) or midazolam.  The non-benzodiazepine sedative zolpidem (Ambien) has demonstrated clear effects on visual BOLD signals with no concomitant change in heart rate or oxygen saturation. The same drug was shown to alter resting connectivity but the accompanying heart rate data were deemed unreliable due to equipment failures. I was unable to find any studies using another common non-benzodiazepine, eszopiclone (Lunesta).

Of those medications that have been considered for possible physiologic confounds, the non-steroidal anti-inflammatory indomethacin was studied a decade ago and found to change BOLD and CBF but not CMRO2 for subjects conducting a simple motor task.

Medicines aimed specifically at increasing a patient's blood flow are of obvious concern for fMRI. Acetazolamide (Diamox) was shown to increase resting cortical perfusion by 20% and decrease primary motor cortical BOLD activation by 35%, whereas the CBF response in primary motor cortex was unchanged in normal volunteers. A similar pattern was observed in patients with steno-occlusive coronary disease. However, another common vasodilator, glyceryl trinitrate, which increases resting cerebral blood volume (CBV), had no demonstrable effect on the BOLD response. The authors suggest that the glyceryl trinitrate affects large arteries only and has negligible effect on the microvascular system driving BOLD.

Blood pressure medications such as angiotensin, or anticoagulants such as warfarin, are presumably not something your subjects will use without telling you; your screening should catch these. Likewise asthma medications. I couldn't find any studies looking at potential physiologic confounds for these medications in any case. But Grichisch et al.investigated the potential effects of nasal insulin on CBF and BOLD and found no difference to a drug-free baseline. (Insulin and blood glucose level is considered again below, under Food.)

Statins (Lipitor, Zocor, Pravachol, etc.) appear to be in widespread and increasing use in western populations. These medicines are designed to reduce cholesterol synthesis in the liver, but there appear to be secondary outcomes with implications for cerebral hemodynamics, such as "...upregulation of endothelial nitric oxide synthase (eNOS) with a subsequent increase in nitric oxide (NO) bioavailability." Xu et al. investigated the hemodynamic consequences of taking atorvastatin (Lipitor) over four months for asymptomatic middle-aged adults. They observed numerous regional changes in BOLD, including both greater and faster regional responses, compared to a group receiving placebo. They observed no change in baseline CBF nor any change in mean transit time (MTT) of cerebral perfusion, suggesting that the regional BOLD changes came about through altered (improved) small vessel vasoreactivity.

Antidepressants, including several selective serotonin re-uptake inhibitors (SSRIs), serotonin-norepinephrine re-uptake inhibitors (SNRIs), and a host of compounds belonging to other four-letter acronyms have been given to subjects in dozens of fMRI studies. But I was unable to find any systematic investigation of concomitant physiology, such as measuring CBF before and after drug administration.

And then there are the so-called "smart drugs" and other brain enhancers such as methylphenidate (Ritalin), amphetamine (Adderall) and their ilk. (If I'm not using the correct collective terms for these compounds, I apologize.) Rao et al. assessed BOLD and perfusion changes for subjects performing a finger tapping task before and after oral methylphenidate and found no changes in neurovascular coupling. Otherwise, I wasn't able to find much literature addressing the potential for physiologic confounds. Marquand et al. looked at CBF changes due to methylphenidate while Nordin et al.assessed CBF changes due to amphetamine, both groups finding regional differences compared to placebo. But any implications for BOLD fMRI are unclear.

Literature coming out of the so-called pharmacological MRI (phMRI) field may provide the best clues for potential confounds to routine fMRI studies, should you have specific reason to be concerned about the use of medicines in your population of interest. Wang et al. reviewed the potential and challenges of using arterial spin labeling (ASL) in phMRI, "...with an emphasis on the methodologies used to control for potentially confounding vascular and systemic effects."

Anesthetics

It's unlikely you'll be surprised by use of anesthetics among your subjects. But there is certainly interest in scanning people who may have had recent surgery, e.g. traumatic brain-injured patients who may have required surgery to treat other injuries or even the TBI itself. You would presumably want to discount any residual neural and physiologic effects of anesthesia. For example, Qui et al. found "Low-dose sevoflurane significantly altered the task-induced CBF-BOLD coupling."Post surgery one expects altered blood pressure, heart rate and breathing rate, so there are plenty of options to measure and control for concomitant physiologic effects.

Foods

It seems that the biggest potential physiological confounds lie with those dietary substances known to alter neural activity, as already reviewed. In this section I'll try to deal with latent effects from things your subjects ingest for other reasons. For sustenance, say. Before that, though, what about the possible effects of restricted nutrition?

There are a couple of fMRI studies on fasting, one that found an effect in a motor task and another that found altered functional connectivity of visual cortex. There are also several reports of altered BOLD signal in response to experimental hypoglycemia (induced via insulin administration), including cognitive, visual stimulation, and sentence comprehension fMRI tasks. All three studies observed regional differences when compared to normal glucose levels. (See Note 2.) Patients with type I diabetes also showed different BOLD activity compared to controls. Additionally, there is one study using CBF and measures of oxidative metabolism (CMRO2) to assess the effects of hypoglycemia. They found that "...metabolism and flow remained coupled. Elementary motor task activation was not associated with any consistent larger activated flows. Thus it remains that although mild hypoglycemia induced an increase in basal flow and metabolism, a similar increase was not seen in task activation."Finally, glucose and fructose infusions were compared to saline and found to alter BOLD signal magnitude in opposite directions.

These are interesting studies, but you're probably more interested in any possible effect of blood glucose level on normal BOLD variability when your subjects aren't fasting intentionally or having their blood sugar levels manipulated artificially. I couldn't find any literature addressing this specific concern, however there is a very recent paper that compared an overnight fast with normal and hyperglycemic conditions. While the paper's abstract makes it sound like there were big changes in BOLD, in the Discussion they write: "These effects are comparatively small, yet may interfere with design sensitivity, when fasting status or blood glucose is not controlled in fMRI experiments."I only skimmed the paper but was left with the impression that the jury is still out on this one. Hyperglycemia, specifically, doesn't seem to be a concern for BOLD. Gruetter et al. observed no significant effects on BOLD for blood glucose of up to 300% of control levels.

Now let's move to what your subjects might be eating before getting a scan. In 2003 it was reported that ingesting lipid - 50 ml of canola oil - decreased BOLD response in a finger tapping task. But the potential effect of triglyceride levels in blood was measured again in 2009 by another group, whereupon no significant effects on BOLD were observed. Why the contradiction? We can only speculate. The authors of the 2009 study offer several plausible explanations in their discussion, and I am inclined to go with the more recent result pending further experimental data.

Then there is the antithesis to the high fat diet: salad. Aamand et al. used a nitrate challenge, "corresponding to the nitrate content of a large plate of salad," to trigger changes in BOLD reactivity without altering the baseline CBF. CBF (with arterial spin labeling) in response to dietary nitrate was also measured by Presley et al.They found no global CBF changes but did observe some regional changes, in frontal lobe. The apparent contradiction with the first report was addressed by Aamand et al. as perhaps due to differences in sensitivity. In Aamand et al. they used only the inferior half of a 32-channel head coil which reduced the SNR in frontal lobe for that study.

Dietary supplements may deserve their own entire section at some point but for now I'll include them as food because they are generally unregulated, unlike FDA-controlled medications. Creatine, usually as the monohydrate, is a popular sports supplement that has been shown to decrease the magnitude of a visual BOLD response by 16%. I didn't find any literature assessing the effects of carnitine, calcium, glycerol, vitamin A (including retinol and beta-carotene), vitamin B (including folic acid and thiamine), vitamin C (ascorbic acid), vitamin D or vitamin E (including tocopherol) on fMRI signals or on cerebral blood flow.

There was also no literature dealing with amino acids generally, but tryptophan has been investigated because of its role as a serotonin precursor. Acute tryptophan depletion, as a way to manipulate serotonin levels, was found to modify regional BOLD responses to a cognitive task, but global effects weren't assessed. I was unable to find any literature investigating dietary glutamate (including as monosodium glutamate) or arginine.

Chronic ingestion of omega-3 fatty acids has been investigated using near-infrared spectroscopy, but only in frontal lobe where changes in both oxyhemoglobin and total hemoglobin were observed. Global measures of CBF weren't feasible. Taken together with prior fMRI results, which reported greater prefrontal BOLD signals in a sustained attention task without a change in task performance, it seems that some sort of vascular changes may be occurring. As with lipid ingestion generally, more work will be needed before we know whether there are general effects of concern for BOLD.

Diurnal factors and sleep

Those of you studying anxiety or the effects of sleep or sleep deprivation are presumably acutely aware of and controlling for potentially confounding physiologic changes in your study designs. The effects of intentional manipulation of cortisol and sleep deficiency on fMRI are well documented. What about the effects in "normal volunteers?"Stress (level of blood cortisol) and amount of sleep (drowsiness) may be factors for the time of day your study is conducted, and they may vary with the populations you sample from. (Ditto for caffeine use!) But I can't find any references specifically addressing these possibilities. There is a now classic study which showed how inter-subject variability is much larger than intra-subject variability, so perhaps it's all baked in. Still, you would want to avoid biasing groups or longitudinal studies by the time of day you scan.

Hormones

I was unable to find any papers dealing with concomitant physiological changes accompanying the menstrual cycle. It seems reasonable to expect that significant effects might be reflected in the heart rate, respiration rate/depth and blood pressure. Direct vasodilatory or vasoconstrictive effects seem unlikely to me, but I'm no endocrinologist. There are papers dealing with performance differences across menstrual cycle, but if these effects repeat they may have a neural basis rather than being physiological artifacts.

Women (and soon men?) going on or off hormonal contraceptives during a study might be a bigger deal. A recent study found a marked change in BOLD responses for women using oral contraception. The effects could be neural or concomitant physiological changes, or a mixture of both.

I couldn't leave this section without mentioning oxytocin. It's quite in vogue. I wasn't able to find any studies that had have investigated systemic physiological effects of this compound as they pertain to BOLD. But giving subjects oxytocin - or any other pharmacological agent come to that - without checking for concomitant physiological changes is taking a leap of faith that all effects are neural. At bare minimum I would expect blood pressure to be recorded with and without the challenge, and then I would want to see heart rate and respiration recorded during fMRI.

Age and disease

These variables should come as no major surprise if you are screening your subjects in any way at all. But in order to be complete and to give you a place to start if you're doing group studies across the lifespan, or comparing disease states to normal volunteers, I offer these papers:

Alterations in the BOLD fMRI signal with aging and disease: a challenge for neuroimaging.

Age-related differences in memory-encoding fMRI responses after accounting for decline in vascular reactivity.

Evidence that neurovascular coupling underlying the BOLD effect increases with age during childhood.

Age-related differences in cerebral blood flow underlie the BOLD fMRI signal in childhood.

Age dependence of hemodynamic response characteristics in human functional magnetic resonance imaging. 

Neural mechanisms of age-related slowing: the ΔCBF/ΔCMRO2 ratio mediates age-differences in BOLD signal and human performance.

Calibrated fMRI during a cognitive Stroop task reveals reduced metabolic response with increasing age.

Visualization of altered neurovascular coupling in chronic stroke patients using multimodal functional MRI.

Acceleration

If your subjects have just returned from a six-month stint on the International Space Station and you want to compare fMRI before and after the flight, you are no doubt fully aware of the many physiological factors to take into account where extended microgravity is concerned. Short duration microgravity - say, the few minutes to be experienced by lucky punters taking Virgin Galactic flights in the not-too-distant future - is probably not much of a concern, unless the flight happens the morning of their fMRI scan.

At the other end of the scale are those people exposing themselves to high acceleration; much greater than 1 g. Presumably these exposures are seconds or at most minutes in duration, but if the exposure occurs frequently or happens in the hours before participating in an fMRI experiment, we should be sure to exclude delayed physiologic changes. To date, though, I can find no reports detailing possible confounding effects of aerobatic plane flights, gymnastics, fairground rides, trampolining and the like. I am assuming these aren't going to be common issues for the vast majority of fMRI studies, but concomitant physiologic changes would be worth consideration if your research happens to involve fighter pilots, race car drivers, etc. Presumably measuring blood pressure, respiration and heart rates would capture the main effects. 





Notes:

1.  The other day someone told me about a subject who had answered negative to a drug screen but who then got very concerned when, on the morning of the scan, she was asked to give a urine sample. (I don't recall why the urine sample was required, and presumably it was stated on the recruitment flier which I can only assume the subject didn't read fully before applying to do the experiment. In any event, it was a vital part of the experiment and not a law-enforcement action!) She apparently returned from the bathroom with a sample of tap water in her specimen tube. The point? People will lie on screening forms if they are trying to make the $20 an hour you're offering. So if your experiment is highly sensitive to illicit drug use then you may want to include the option (or the threat?) of a urine test to help filter those people out.

2.  Hypoglycemia was also one of the conditions used (in rats) by Ogawa et al.in their preliminary demonstration of the BOLD effect. From the abstract of their 1990 paper: "This blood oxygenation level-dependent (BOLD) contrast follows blood oxygen changes induced by anesthetics, by insulin-induced hypoglycemia, and by inhaled gas mixtures that alter metabolic demand or blood flow."So we're not exactly covering new ground here, are we?

This post updates the checklist that was presented back in January, 2013. The updated checklist is denoted Version 1.2. The main update is to include reporting for simultaneous multi-slice (SMS) (a.k.a. multi-band, MB) EPI.

Explanatory notes for parameter names appear in the lower portion of the post. Note that the present checklist was devised by considering typical fMRI experiments conducted on 1.5 T and 3 T scanners but the list should work reasonably well for 7 T sequences, too.

Please keep the comments and feedback coming. This is an ongoing, iterative process.




Release notes for Version 1.2

All changes from Version 1.1 have been highlighted in yellow, both on the list PDF and on the explanatory notes (below).

1. The "Spatial Encoding" parameter categories have been renamed "In-Plane Spatial Encoding" to better differentiate in-plane acceleration (e.g. GRAPPA) from slice dimension acceleration (SMS/MB).

2. When using slice dimension acceleration (i.e. SMS/MB), certain parameters that are listed as Supplemental for other EPI variants should be considered Essential. Specifically, it is suggested to report:

Matrix coil mode
Coil combination method

All the In-Plane Spatial Encoding parameters in the Supplemental category should be considered because there is a tendency to use SMS/MB to attain high spatial resolution, requiring long readout echo trains that can have higher distortion than found in typical EPI scans.

The In-plane reconstructed matrix parameter should be reported whenever partial Fourier sampling is used, as it often is for SMS/MB EPI.

All the RF & Contrast parameters in the Supplemental category should be reported because the shape, duration and amplitude of the excitation RF pulse are all integral components of the acceleration method.

The Shim routine should be reported if a non-standard shim is performed before SMS/MB EPI.

3. Pre-scan normalization has been added to the Supplemental section of RF & Contrast parameters. Large array coils produce strong receive field heterogeneity and the use of pre-scan normalization may improve the performance of post hoc motion correction.



Explanatory notes


Essential - Scanner

Magnetic field strength: In lieu of magnetic field strength the scanner operating frequency (in MHz) might be considered acceptable. I'm assuming we're all doing 1-H fMRI. If not, chances are the methods section's going to be very detailed anyway.


Essential - Hardware Options

Rx coil type: For standard coils provided by the scanner vendor a simple description consisting of the number of independent elements or channels should suffice, e.g. a 12-channel phased array coil, a 16-leg birdcage coil. Custom or third-party coils might warrant more detailed information, including the manufacturer. Most head-sized coils are Rx-only these days, but specifying Rx-only doesn't hurt if there could be any ambiguity, e.g. a birdcage coil could quite easily be Tx/Rx.


Essential - In-Plane Spatial Encoding

Pulse sequence type: A generic name/descriptor is preferred, e.g. single-shot EPI, or spiral in/out.

Number of shots (if > 1): Multi-shot EPI isn't a common pulse sequence; single-shot EPI is by far the most common variant, even if acquired using parallel imaging acceleration. I include this option to reinforce the importance of reporting the spatial encoding accurately.

PE acceleration factor (if > 1): This is usually called the acceleration factor, R in the literature. Vendors use their own notation, e.g. iPAT factor, SENSE factor, etc.

PE acceleration type (if > 1): It seems that most vendors use the generic, or published, names for parallel imaging methods such as SENSE, mSENSE and GRAPPA. I would think that trade names would also be acceptable provided that the actual (published) method can be deciphered from the scanner vendor and scanner type fields. But generic names/acroynms are to be preferred.

PE partial Fourier scheme (if used): Convention suggests listing the acquired portion/fraction of k-space rather than the omitted fraction. Any fraction that makes sense could be used, e.g. 6/8 or 48/64 are clearly equivalent.


Essential - Spatial Parameters

In-plane matrix: This should be the acquired matrix. If partial Fourier is used then I would suggest reporting the corresponding full k-space matrix and giving the partial Fourier scheme as listed above. But I wouldn't object to you reporting that you acquired a 64x48 partial Fourier matrix and later reported the reconstructed matrix size as 64x64. So long as everything is consistent it's all interpretable by a reader. (But see also the In-plane reconstructed matrix parameter in the Supplemental section.)

In-plane inline filtering (if any): Non-experts may be unaware that filtering might be applied to their "raw" images before they come off the scanner. It's imperative to check and report whether any spatial smoothing was applied on the scanner as well as during any "pre-processing" steps subsequent to porting the time series data offline.

Slice thickness and Inter-slice gap: For now I would use the numbers reported by the scanner, even though there may be some small variation across scanner vendors and across pulse sequences. For example, some vendors may use full width at half slice height while others may use a definition of width at or near the base, there may be pulse sequence options for RF excitation pulse shape and duration, etc. I see these details as secondary to the essential reporting improvements we're aiming for.

Slice acquisition order:Interleaved or contiguous would be sufficient, although explicit descending (e.g. head-to-foot) or ascending (e.g. foot-to-head) options for contiguous slices would be acceptable, too. Presumably, the subsequent use of slice timing correction will be reported under the post-processing steps (where most fMRIers call these "pre-processing" because they are applied before the statistical analysis).


Essential - Timing Parameters

TR: For single-shot EPI with no sparse sampling delay the TR also happens to be the acquisition time per volume of data. But if sparse sampling or multiple shot acquisition is being used then the TR should be clearly reported relative to these options. The conventional definition of TR is the time between successive RF excitations of the same slice. Thus, by this definition the reported TR would include any sparse sampling delay, but it would specify the time of each separate shot in a multi-shot acquisition and the per volume acquisition time would become TR x nshots.

No. of volumes in time series: Dummy scans (not saved to disk) should be reported separately. Likewise, the use or rejection of the first n volumes for task-related reasons, e.g. to allow a subject to acclimatize to the scanner sounds, should also be reported separately in the post-processing segment of the experiment. In this field we are only interested in the total quantity of data available to the experimenter.

No. of averages/volume (if > 1): I don't think I've ever seen anyone do anything but one average per TR for single-shot EPI/spiral (unless they've screwed something up) and I can't think of a reason why someone would want to do it for fMRI. But, if it happens for some reason then it's really, really important to specify it.


Essential - RF & Contrast

Fat suppression scheme: It's sufficient to state that fat saturation or fat suppression was used, for example. Further details aren't required unless the scheme was non-standard, e.g. a custom spatial-spectral excitation scheme.


Essential - Slice Acceleration

SMS/MB acceleration factor:  The number of slices excited simultaneously, equivalent to the spatial acceleration in the slice dimension. The total number of slices (Number of slices parameter) must be divisible by the SMS/MB acceleration factor. Note that the slice dimension acceleration (this parameter) is independent of any in-plane acceleration (e.g. SENSE, GRAPPA) and should be reported independently. For example, GRAPPA with R=2 should be reported as suggested in the In-Plane Spatial Encoding section, while slice acceleration should be reported here. It is not acceptable to combine the parameter reporting as one overall acceleration because the practical consequences of each are very different.

Sequence/recon name &/or version:  Unlike many single-shot EPI sequences which are product, most SMS/MB sequences are being actively developed and change frequently. It is imperative to report version numbers for both the acquisition and, if they aren’t coupled, the reconstruction. If reconstruction is performed offline, report the method used, and any version number, in full.


Essential - Customization

Sparse sampling delay (if used): Sometimes called "Delay in TR" on the scanner interface. Used most often for auditory stimulus or auditory response fMRI.

Prospective motion correction scheme (if used): PACE is one commercial option. These schemes fundamentally change the nature of the time series data that is available for subsequent processing and should be distinguished from retrospective (post-processing) corrections, e.g. affine registration such as MCFLIRT in FSL. It is also critical to know the difference between motion correction options on your scanner. On a Siemens Trio running VB15 or VB17, for instance, selecting the MoCo option enables PACE and a post hoc motion correction algorithm if you are using the sequence, ep2d_pace whereas only the post hoc motion correction algorithm - no PACE - is applied if you are using the ep2d_bold sequence. There's more detailed information on these options in my user training guide/FAQ.

Cardiac gating (if used): This isn't a common procedure for fMRI, and recording of cardiac information, e.g. using a pulse oximeter, is radically different from controlling the scanner acquisition via the subject's physiology. The recording of physiological information doesn't usually alter the MRI data acquisition, whereas gating does. Reporting of physio information is tangential to the reporting structure here, but if you are recording (and using) physio data then presumably you will report it accordingly somewhere in the experiment description.


Supplemental - Hardware Options

Gradient set type: It should be possible to infer the gradient coil from the scanner model. If not, e.g. because of a custom upgrade or use of a gradient insert set, then the specifications of the actual gradient coil should be reported independently.

Tx coil type (if non-standard): It should be possible to infer the Tx coil from the scanner model. If not, e.g. because of a custom upgrade or use of a combined Tx/Rx coil, then the new Tx coil should be reported independently. I would also advocate including the drive system used if the coil is used in anything but the typical quadrature mode.

Matrix coil mode (if used): There are typically default modes set on the scanner when one is using un-accelerated or accelerated (e.g. GRAPPA, SENSE) imaging.  If a non-standard coil element combination is used, e.g. acquisition of individual coil elements followed by an offline reconstruction using custom software, then that should be stated.

Coil combination method: Almost all fMRI studies using phased-array coils use root-sum-of-squares (rSOS) combination, but other methods exist. The image reconstruction is changed by the coil combination method (as for the matrix coil mode above), so anything non-standard should be reported.


Supplemental - In-Plane Spatial Encoding

PE direction: If you've shown any examples of EPI in your paper then the PE direction can usually be determined from the image. If N/2 ghosts or distortions aren't obvious, however, then it's rather important that the phase encode direction is stated, in concert with the readout echo spacing, so that a reader can infer your spatial accuracy.

Phase oversampling (if used): There's no reason to use phase oversampling for EPI - you're wasting acquisition time - but if you are using it for some reason then it should be reported consistent with the acquired matrix, acquired FOV, echo spacing and associated parameters.

Read gradient bandwith: Not an intrinsically useful parameter on its own, it does have value if reported in conjunction with the echo spacing. (An alternative to this single parameter would be the read gradient strength (in mT/m) and the digitizer bandwidth (in kHz).)

Readout echo spacing: Rarely reported but really useful! This number, in conjunction with the FOV and acquired matrix size, allows a reader to estimate the likely distortion in the phase encode direction.

Pulse sequence name: Could be invaluable for someone wanting to replicate a study. There may be multiple similar pulse sequences available, all capable of attaining the specifications given, but it is entirely feasible that only one of the sequences has a particular quirk in it!

k-space scheme: Readers will assume linear (monotonic) k-space steps in the PE direction unless indicated to the contrary. Centric ordering or other atypical schemes should be indicated, especially in concert with multiple shots if the Number of shots parameter is greater than one.

Read gradient strength: Could be useful in conjunction with information about the ramp sampling percentage and echo spacing time, otherwise probably of limited value to most readers.

(Ramp sampling percentage:) Ramp sampling can increase the N/2 ghost level considerably if there is appreciable gradient and digitization (data readout) mismatch. But determining the percentage of readout data points that are acquired on the flat vs. the ramp portions of each readout gradient episode can be involved. And for routine studies, as opposed to method development studies, there's probably not a whole lot of value here. Maybe remove it?

(Ghost correction method:) N/2 ghost correction usually happens invisibly to the user, but there are some options becoming available, especially useful for large array coils (e.g. 32-channel coils) where there may be local instabilities with some ghost correction methods. If known, and if non-standard, then it would be nice to report. But perhaps more overkill for fMRI methods?

PE partial Fourier reconstruction method: If the scanner offers more than one reconstruction option then the chosen option should be reported.


Supplemental - Spatial Parameters

In-plane resolution: This field is redundant if the reconstructed matrix (In-plane matrix parameter) and FOV are reported, but I for one wouldn't object to seeing the nominal in-plane pixel size given anyway. It may make the paper a faster read. Probably not worth arguing about. (Cue extensive debate...!)

In-plane reconstructed matrix: This is for reporting of zero filling (beyond the default zero filling that may have been done for a partial Fourier acquisition) to a larger matrix than acquired, prior to 2D FT. There may be modeling issues associated with the number of voxels in the image, not least of which is the size of the data set to be manipulated! It could save someone a lot of angst if she knows what you did to the data prior to uploading it to the public database.


Supplemental - Timing Parameters

No. of dummy scans: This is the number of dummy scans used to establish the T1 steady state. Many fMRI experiments also use acquired volumes subsequent to the (unsaved) dummy scans for neuroscientific reasons, e.g. to allow definition of a BOLD baseline or adjustment to the scanner noise. These two parameters should be reported separately.


Supplemental - RF & Contrast

Excitation RF pulse shape and Excitation RF pulse duration:  Not critical for standard pulse sequences on commercial scanners, but if atypical values are set in order to achieve very thin slices, for example, then reporting these parameters may be benenficial. These two parameters are essential, however, when reporting SMS/MB EPI.

Pre-scan normalization: Typically achieved with a low-resolution gradient echo scan, but the actual process may be invisible to the user. Instead, activating pre-scan normalization may simply require selection of the appropriate option. (This is the case on Siemens scanners.)


Supplemental - Slice Acceleration

SMS/MB reconstruction type:  Report details of the reconstruction type if possible, e.g. the kernel size for a GRAPPA-style reconstruction.

FOV shift:  The sequence may use blipped controlled aliasing along the slice direction to assist in the separation of the simultaneous slices. The blips are set to produce a fixed partial FOV shift, typically FOV/2, FOV/3 or FOV/4. The FOV shift may be user-controlled or set to a fixed default.

SIR/SER factor:  If simultaneous image (or echo) refocusing is used in addition to SMS/MB (as in Feinberg et al., PLoS One, 2010), report the SIR/SER acceleration factor as well as the SMS/MB acceleration. 


Supplemental - Customization

Image reconstruction type: Unless specified to the contrary, most readers will assume that magnitude images were taken from the 2D Fourier transform that yielded each individual EPI. If you use complex data - magnitude and phase - then that option should be specified along with the particular processing pipeline used to accommodate the atypical data type.

Shim routine: If manual shimming or an advanced phase map shimming routine is used, especially to improve the magnetic field over restricted brain volumes, then this information should be reported.

Receiver gain: Most scanners use some form of autogain to ensure that the dynamic range at the receiver is acceptable. If manual control over receiver gain is an option and is used then it should be reported because a mis-set gain could lead to artifacts that aren't typically seen in EPI, and a reader could subsequently attribute certain image features to other artifact sources.





Abbreviations:

FOV - Field-of-view
MB - Multi-band
N/2 - Half-FOV (for Nyquist ghosts)
PE - Phase encode
Rx - Radiofrequency (RF) receiver
SER - Simultaneous echo refocusing
SIR - Simultaeous image refocusing
SMS - Simultaneous multi-slice
Tx - Radiofrequency (RF) transmitter
TE - Echo time
TR - Repetition time

Crazy Scientist sent me a link to a paper, "Neuroimaging after mild traumatic brain injury: Review and meta-analysis," (doi:10.1016/j.nicl.2013.12.009) and it prompted me to do something with a brief review I wrote this time last year as a way to plan some research activities on MRI of mild traumatic brain injury (mTBI). By a remarkable coincidence the review paper was made available online four whole days before I completed my own review. I've yet to read the published review so I can't yet tell if I wasted my time. In any event, the document I wrote was for internal consumption (for my collaborators, and to clarify my own thoughts) and was never designed to become a public document. But since our research direction changed mid-year I figured I might as well stick my review out there in case anyone can make use of it.

The title of my document is the same as the title of this post. You will find the contents pasted below, or if you prefer you can download a PDF from this Dropbox link. I have quickly re-read it to check for major bloopers, and I've added a couple of update notes highlighted in yellow. There may well be some direct copy-paste of parts of a few of the papers I reviewed, especially those with heavy neuroradiology content where I am generally a long way out of my depth and would prefer to accept charges of plagiarism than get the medical terminology wrong!

For the record, we are still interested in mTBI but the logistics of studying acute brain injury in a non-hospital setting, using a home-made machine (the ULFMRI) sitting in a second basement lab in a physics department, made it all too hard to pursue right now. We have shifted instead to studying chronic conditions where we have a fighting chance of getting a few people scanned in our unorthodox facilities.

Open access online science publisher The Winnower has made a huge leap in scientific publishing and now offers bloggers a way to assign a digital object identifier (DOI) to any post uploaded to their site. Once there, the post can be reviewed, etc. just like any other online paper. They will also be archiving soon via CLOCKSS. (See Note 1.) With a DOI plus archiving it means that a post (and any reviews) should be traceable in perpetuity. Very good developments!

I had the privilege of helping Josh, The Winnower's industrious founder and administrator, create a new publication category for Neuroimaging and then used my post on physiologic confounds as the first test case. The post now has its own DOI (DOI: 10.15200/winn.142919.97862 ) should you prefer that to a URL. I submitted via a Word document because I've found that Blogger will occasionally hide irrelevant HTML code that needs to be edited out by hand. This can be a problem if you're in the habit, as I am, of copy-pasting text (e.g. quotes from papers) into a blog post. There is a new facility that will submit directly from a blog but in my first test (using Blogger) there were some major formatting issues. Josh informs me that he will be adding a facility for the Blogger API eventually, but using Word as an intermediate step gives me a chance to clean up links to references and that sort of thing. I have plans to submit many other posts to The Winnower so please let me know either in the comments below, in a review on the The Winnower version of the post or via Twitter whether you encounter problems, have suggestions for improvements, etc. Also, do please consider submitting your own blog posts to The Winnower. Let's build the Neuroimaging category!

I was around for and involved in the nascent Web of the early '90s (see Note 2) and I distinctly remember NCSA's Mosaic, the Planet Earth Home Page virtual library, UC Irvine's online bookshop, and the Cambridge coffee pot. But the development of open, online publishing supported by social media like blogs and Twitter, as well as post-publication peer review (PubPeer and PubMed Commons), feels like a true revolution for science. I may be wrong but it feels like we will look back at the current period as a major change in the way we interact. It only took us two decades. Now, however, The Winnower is contributing disproportionately to our future. Thank you.



Notes:

1.  I already archive my important blog posts at The Internet Archive's Wayback Machine. In addition to providing an invaluable service, the Wayback Machine is also a wonderful way to get nostalgic and kill a few hours :-/

2.  I had an online poster at NMR Poster95, the first e-poster meeting of folks doing NMR and MRI. The front page of the poster got archived but the clickable poster itself is now defunct, I'm afraid. The wonderful Internet Archive does have other pages from my website at the time, however. They began archiving in 1996, and the earliest copy of my old website dates from January, 1997. Thank you, Internet Archive! There is also a summary of the first two NMR e-poster conferences in this paper from 1997.

So I have a memory like a sieve except that it's profoundly less useful in the kitchen. And because I know from painful experience that anything I don't document never happened, I am going to help myself and you by creating in real time a tutorial to upload blog posts from Blogger to The Winnower, should you be so inclined. Why do it? DOI is one reason.

Those of you who were smart enough to begin your blog's existence on Wordpress can use a fancy plugin for your API. Those of us who now have too much inertia on Blogger to relocate must do a little more work and use some intermediate steps, but it really isn't that hard. What's more, the intermediate steps offer an opportunity for proofreading and fine-tuning that you might like to do anyway. Let's do it!


Initial steps

First up, I assume you have created an account on The Winnower.  Then select the SUBMIT option in the top menu. At that point the step-by-step instructions can be found in the link in the top-right corner of the page, Submit Your Paper:



You'll get a new window pop up; skip on down to the Word instructions:


Here you will want to download the Word template and also read through all the formatting guidelines before moving on:



So now let's shift to the template document just downloaded. Aha! This looks familiar! I couldn't for the life of me recall how I'd managed to have on my laptop the neatly formatted docx file of an earlier post I'd uploaded to The Winnower last December. The memories are beginning to flow. What a relief!


Okay, so now I need to get myself offline for a bit and copy-paste today's upload into the template. I don't recall hitting any major obstacles previously but if I do I shall be sure to report them here in a mo.

I'm back. The first thing I had to do was decide the source of the material I'm going to copy-paste. I keep PDF copies of all my posts so I began by trying that, but it's not ideal because of all the extraneous words in the menus, etc. when you print to PDF from a browser. Makes it hard to get a clean text selection over anything more than a paragraph. Thus, I elected to go back into Blogger API and open the post I'm uploading, as if I am going to edit that post. The Blogger API is pretty simple although it has a habit of hiding irrelevant HTML if you've pasted text from formatted sources. In any event, the Blogger API makes it convenient to separately copy-paste the title and post body. I edited the author and address information by hand.

The abstract is my first opportunity to revise what's in the blog post for The Winnower. The post I'm uploading had two prior incarnations: a preliminary draft with a call for input, followed by a first completed version. An introduction to the post resides only in the first version but it needs to be updated to reflect the evolution in subsequent posts. Okay, then. Snip, snip, splat, done. A new abstract is created!



References!

Here's an important point. Notice what I did and didn't do with the Poldrack reference in my abstract. * The first post I uploaded to The Winnower contained a bazillion references. It was a literature review! I went back and forth with Josh about the citation recommendations in his guidelines. I was beta-testing the upload procedure so I tried to play along, except that the way I'd used simple hyperlinks (mostly to PubMed IDs) in the blog post meant that I would have many days' work to conform to the guidelines. Once the blog was uploaded and all the links worked I decided that it was good enough to proceed with the links alone. Check it out, see if you agree. I am assuming that PubMed and the NIH will be around a while so I'm counting on PMIDs to be resolvable indefinitely. Of course, I see the irony in not using a DOI instead of the PMID here given one of the goals of this upload, but I tend to use PMIDs a lot more than DOIs to this point. If in doubt, do both. Otherwise, I'm going to give my personal opinion that it is better to have got your post uploaded with PMIDs, links or whatever you've got than to procrastinate for wont of conforming to some standard. Sure, DOIs and full citations for every reference would be lovely, but we're busy here! The Winnower papers work perfectly well with hyperlinks so provided your links have some survivability to them I think you're good to go. **

* I just noticed that links aren't given differential formatting automatically in the abstract of The Winnower articles, making them hard to see. But you can mouse-over them and they reveal themselves. There are four links in the abstract to my first uploaded post. See if you can find them. In today's test I shall add bold, italics and underline to the abstract template and see if I can emphasize the links...

** Links in the manuscript body (that is, anything outside of the abstract) format automatically to reveal themselves, so just insert links as usual in either the Blogger API or in Word and you're good to go.


Now it's time to copy-paste the body of the post between the [manuscript] anchors of the template. Pretty easy stuff, except that again I'm going to take the opportunity to tweak some things. I'm going to move the abbreviations up to the top, and I need to address the issue of highlighting and other formatting that resides in the blog post.

I'm busily copy-pasting from the Blogger API into the template. I'm grabbing entire chunks of text, figures 'n all. I simply select the "Match Destination Formatting" option that Word gives me whenever a new pile of stuff is dropped wholesale into the template. Some tabs are vanishing so there's a little bit of work comparing the format of the published blog post to the template draft, but so far so good. Though it should be noted that I expect a lot of my reformatted text to appear differently once it's converted to HTML and uploaded to The Winnower, so I'm not spending very much time making my document purdy.

Figures and tables

I'm also being pretty cavalier with figures and tables. In this post the only figures are actually tables of variables. The table extends slightly more than one page. Rather than spend goodness knows how much time creating a new table or making everything fit on a single page, I'm again going for expediency. So it's Word > print to PDF > export as PNG > insert into template. Sorry, Josh. We're online here, so all that really matters is that the final document is readable. If I've wasted a third of someone's screen then I apologize. I refer you to my earlier opinion regarding overcoming procrastination.

The upload

I'm nearly there! The template document is complete and I've double-checked that the abstract lies between the [abstract] anchors and everything else resides within the [manuscript] anchors. It's upload time.

The next step is to export an .htm version out of Word. At this point my document is called checklist_to_Winnower.docx rather than word_template.docx. I'm going to assume that The Winnower's servers like to discriminate between uploads. That said, I don't actually think it matters what you call your .docx or the subsequent .htm file. So I'm just selecting "Save as Web Page..." per the instructions: ***



*** The option to "Save entire file into HTML" creates a number of .xml that aren't needed for the upload. Thus, you can opt for "Save only display information into HTML" when exporting from Word.  

Note the requirement to select the "Web Options..." and then ensure you have the Unicode (UTF-8) option selected:


Word immediately opens the .htm file for me but as a skeptic I will quickly review the file in my browser. The first part looks okay except for a deliberate mistake :-/ Anyone spot it?



Yes, that's right. My post is presently entitled "Windows Template." So just a few moments while I correct that oversight... Done. I have to export to .htm, of course, so I'll spend a few minutes checking the rest of the export first. I rather like the idea of enforced proofreading. This isn't quite enforced but the hassle of re-exporting certainly encourages it here. Hands up who has reviewed a paper where one of your earliest thoughts is along the lines of "WTF? Do they expect me to proofread it then review it? Can I be a co-author?" Spend the time, people. 99 percent done, halfway there.

Okay, so I've scan-read the .htm file (I don't take my own advice) and I'm ready to proceed. I notice a new folder containing the images and some .xml files. The instructions tell me I'll be needing the image files, at least, during the upload. Here goes.



The first page of the upload is self-explanatory, but you do need to come up with up to five keywords. Make them count. Somewhat usefully, the interface shows you all the keywords that people have used before although they're not categorized.

The next page asks me to upload my document, or I could perhaps use a URL to the .htm file residing in my browser. I haven't tried the latter before so I'll use the DOC button, as I think I did last December. Then I just have to find and upload the .htm file and the three image files in the post:




Having checked the two radio buttons and selected PROCESS UPLOADED FILES you will definitely want to preview your paper. My draft is well spaced. Too well spaced, in fact. A few diagnostics shows that a new line after an anchor, a header, a sub-header, a figure or a paragraph is sufficient to insert a half line space. A blank line results in 1.5 lines in the uploaded version. So I'm going to go back and remove the blank lines between my paragraphs, etc. in the Word document. It makes the Word version look crap but I'm confident the formatting will work properly once The Winnower gets a hold of it. This is what it now looks like as .htm with blank lines removed:


It looks clunky as .htm (and in Word) but once uploaded the formatting gets sorted out quite nicely.

I've previewed my paper and it's good enough. I notice the links in the abstract are displaying bold and italicized but not underlined (two out of three ain't bad!) which discriminates them nicely from the plain text:


While it's not perfect, the theme of this post is to "git r' dun" so I'm going to resist the temptation to get anal at this late stage and instead select the PROCESS UPLOADED FILES button to release the beast.

Oh, this is cool. On the third of the four upload pages you are given the option to provide "additional assets" to your paper. I was expecting to be asked for my credit card and CVV code. But no! It's a place I can drop a PDF version of the checklist, to compliment the one that already resides on Dropbox. Nice.

All that remains to do is notify anyone I want to review the paper - no thanks, I'll keep it a secret ;-) - and.... Voila! Success!!!!



The new paper is now live right here. I've checked that the links and the downloadable PDF work properly but I'll leave any more strenuous proofreading and corrections for a later date. Likewise, I'll wait on assigning a DOI because I don't see the button that was there when I used it a month ago on another post. For now, then, it's a wrap.

The latest version of the fMRI acquisition checklist is now available at The Winnower via this link. It can also be located/cited using DOI: 10.15200/winn.143191.17127 .

• The “Pre-scan normalization” parameter has been renamed “Signal intensitycorrection/normalization” to broaden its scope.
• Reviewed and revised explanatory notes.
• New parameters: RO partial Fourier scheme, Number of echoes, Saturation bands, Gradient non-linearity correction, Z-shim gradient correction, On-resonance adjustment.

The current list should work now for most simultaneous multislice (SMS) EPI, multi-echo EPI (or spiral) and 7 T experiments, but additional emphasis will be placed on these advanced methods in the next round. Please let me know of parameters you'd like included. The next planned update will include as far as possible the vendor-specific nomenclature for each parameter. I anticipate a late 2015 release date. There are no plans yet to do a machine-readable version of the checklist, as originally discussed, but that is only because nobody has been asking for it. Please get in touch if this is something you're interested in.

One of the delightful aspects of running an imaging facility is the sheer variety of projects coming through the door. Late last year my boss told me he'd been discussing with a group from Emory University about doing fMRI on trained dogs at our center. I'll confess to receiving the suggestion unenthusiastically, if only because I envisioned a mass of bureaucracy followed by a head-on logistical collision between the dog group and the dozens of human users. Activity at our center oscillates between hectic and frenetic, depending on the day. But, as it turned out I needn't have worried. The bureaucracy was handled admirably by the Emory folks while the logistical issues simply failed to materialize because of the professionalism of the dog fMRI team. It's been an enjoyable experience. And there are dogs. Many boisterous, happy, playful yet exceedingly well-trained dogs. Like these:



Motivation

Greg Berns (on Twitter here) at Emory decided a few years ago to do something that few thought was possible: to scan dogs with fMRI while they are awake and behaving, using nothing more than the same tactics as we prefer to use for human MRI subjects, i.e. training. No sedatives, no anesthetics. Nothing but some good old-fashioned familiarization.

But why try to do dog fMRI at all? In Greg's case the initial question was what fMRI might tell him about his own pet dog's brain. A simple example: was his dog excited to see Greg come home from work because it was the prelude to dinner - a conditioned response - or was the dog genuinely pleased to see Greg, the person, and dinner was a nice bonus thank you very much? I'll let Greg take it from here, courtesy of a recent TEDx talk:



Greg's initial dog fMRI project also featured in a segment broadcast on CBS's 60 Minutes in October, 2014.

As mentioned in the TEDx video, the project that Greg wanted to run at my center uses service dogs-in-training. All the dogs are from Canine Companions for Independence (CCI), based in Santa Rosa, CA. Ignoring Bay Area traffic our location is sufficiently convenient to conduct scans on these dogs. We also happen to have the same scanner that Greg has at Emory.


Training

In human fMRI research the vast majority of the time and effort goes into the data processing and statistics. For dog fMRI, on the other hand, there is a truly massive front end load to train the subjects to get them into the scanner in the first place. It's a huge, intricate, time-consuming undertaking with each dog going through up to three months of MRI-specific acclimatization to the sights, sounds and feel of the scanning environment. Learning the task they're to perform is trivial by comparison. But it's far from a burden. As service dogs-in-training these animals spend many hours of every day gleefully learning to locate correctly various objects left laying around the CCI campus, or how to open a door with a rope, or how to react to a fire alarm versus a door bell, or any number of other vital activities they will eventually perform out in the real world.

I got a tour of the mock scanner at CCI-Santa Rosa with Kerinne (on the right, below) and Erin, two of the awesome CCI dog trainers. You can also see Fritz, our volunteer for the day, in his kennel behind Erin's right elbow.



As with any human MRI the first task is safety. For dogs it's extra important to protect their sensitive hearing. Unlike humans, dogs tend to shake earplugs out of their ears unless the devices are literally strapped in, so the dogs are fitted with a sports wrap to hold the plugs in place:



Over the course of several weeks the dogs are trained to place their heads in a styrofoam muzzle rest positioned inside a wooden ring mimicking an RF receiver coil. They also learn a task, signaled by the trainer, that tells them when they can expect to get a small food reward. Eventually, the dogs learn to climb a set of steps and lie down with their muzzles on the chin rest while they lie still in a mock MRI, which beeps and clicks using recordings from the real thing. Unlike the real thing, however, it is perfectly safe for me to record videos from all angles and up close. Here's Fritz during a training session:





Scan day

The dogs are trained in groups, typically 8-10 at a time, to be scanned over a single weekend. On scan day the dogs are chauffeured over to Berkeley with their trainers and assistants. Luckily, not only is Berkeley a dog-friendly campus but it also has many open spaces for the CCI handlers to exercise the dogs while they wait. As you might imagine, a day trip out with their friends can get dogs pretty excited so the CCI staff have a bonus challenge to keep their charges focused. Nobody said hard work couldn't also be fun!

The setup at the scanner matches very closely that in the mock scanner, except for the addition of mats and covers to protect the magnet and the scanner suite from bits of dog treat and dog hair. We don't get too much of the latter because all the subjects get a bath the day before. Each dog is introduced to the scanner suite for a quick familiarization and as soon as the humans are ready (most delays are for the humans!) the dog goes through its now familiar routine.

Interestingly, none of the dogs scanned so far seems to be bothered by the presence of the magnetic field. Whether they sense and ignore it or cannot sense it we don't know. In any event their reactions to the magnetic field are not appreciably different than the average human subject. The advantage for the dogs, as for humans who have been trained in a mock scanner, is that they are acclimatized to the dark, confining tube that makes lots of noise. The biggest distraction seems to be the novel environment they've just experienced - the car ride, the Berkeley campus, new humans around the imaging center - but going from play mode to work mode is a quality specifically selected for in service dogs.

After trying different coil options Greg settled upon the standard Siemens 2-channel human neck coil for the dogs. It's sufficiently large and open that the dogs can lie in the sphinx position to get their heads inside, with front paws either side, and the handler can stand behind the magnet to signal the behavioral task. The neck coil generates decent images. Here's a screenshot of an anatomical scan on the left and a mosaic of EPIs on the right:




There's plenty of contrast while the spatial resolution and SNR aren't too bad for a brain about the size of a lemon. Dogs have thick skulls as well as large muscles around the head. These dogs are either retrievers or Labradors about a year old and are all fit and healthy so subcutaneous lipid signal tends to be similar to what I'm used to seeing with human heads. Fat suppression is used for EPI just as for human fMRI. Other parameters are described in this recent PeerJ paper:

Functional scans used a single-shot echo-planar imaging (EPI) sequence to acquire volumes of 24 sequential 3 mm slices with a 10% gap (TE = 28 ms, TR = 1,400 ms, flip angle = 70°, 64 × 64 matrix, 3 mm in-plane voxel size, FOV = 192 mm). Slices were oriented dorsally to the dog’s brain (coronal to the magnet, as, in the sphinx position, the dogs’ heads were positioned 90° from the usual human orientation) with the phase-encoding direction right-to-left. Sequential slices were used to minimize between-plane offsets from participant movement, and the 10% slice gap minimized the crosstalk that can occur with sequential scan sequences.

If you're interested in the entire scan protocol, including some of the earlier approaches that have been improved upon in the current experiments, there's more information in this PLOS ONE paper from 2013.


Dealing with motion

Motion is the arch enemy of fMRI whatever the species. As already mentioned, the best approach is to have compliant subjects. Greg's muzzle/chin rest design serves twin purposes. It allows the dogs to relax in a position where the head is reasonably well restrained left-right and head-tail, and it is a mark for the dogs to return to whenever they move to get a reward. As with human fMRI, watching the inline display as the EPIs are acquired is highly informative. It's not uncommon for a dog to be able to go 60 seconds with motion effects visible only at the level of the N/2 ghosts, as for a motivated human subject. And as for returning to their marks, it has to be seen to be believed. I once spent several minutes being highly amused by the ability of one dog to get a treat, swallow it and then return to the same position such that a triangle of small blood vessels ended up in precisely the same position in the same EPI slice. I had my finger on the screen as a reference. Except for the contrast changes due to blood flow in those vessels there were no other signs of life!

The EPI is run continuously while the dog does the task and gets treats along the way. The entire run is thus contaminated by numerous intended movements - for rewards - as well as the occasional unintended movement - most often a dog licking/swallowing a few seconds after returning to its mark. An observer notes the times in the scan the dog is being treated, to assist in the elimination of motion in post processing, but as with human fMRI the effects of gross head motion tend to be unambiguous whether they're intended or not. Unlike humans, however, these dogs tend not to fidget with the rest of their bodies, not even wagging their tails when they are on task. And they don't fall asleep, either.

I'm not involved in the processing so I can't comment on how well or how poorly one can chop out motion-contaminated EPI frames and splice together the remainder for analysis. A major difference with most task-based fMRI is that Greg's team simply discard the majority of frames where motion occurs. The trick, then, is to decide the cutoff for discarding data. This is the procedure outlined in Greg's most recent PeerJ paper:

Because dogs moved between trials (and when rewarded), aggressive censoring was carried out, relying on a combination of outlier voxels in terms of signal intensity and estimated motion. Censored files were inspected visually to be certain that bad volumes (e.g., when the dog’s head was out of the scanner) were not included. The majority of censored volumes followed the consumption of food. On average, 51% of total EPI volumes were retained for each subject (ranging from 38% to 60%).

Other than aggressive censoring, the time series are also motion-corrected using a 6-parameter affine registration and the motion traces are used as regressors in the statistical analysis. Standard stuff. Whether this is better or worse than what is commonly done for human fMRI I'm not qualified to say. What I do like is the expectation that a lot of data will be discarded and this is budgeted for in the acquisition. With most human fMRI the motion issue tends to be considered post hoc and rarely do people expect to throw out a lot of data.




Future developments

What could be done to improve dog fMRI? I'd like to test our ability to measure head movement with a peripheral device such as a pressure sensor placed on the chin rest. We have the equipment and we could probably incorporate the sensor, which in our case is a circular disk less than 2 cm diameter, into the chin rest in such a way that the dog doesn't even notice it. We record movement (usually chest movement for humans) using BIOPAC equipment which also records TTL signals sent once per TR during the EPI time series, so synchronization is trivial. Perhaps we could use such a measure to define automatically those segments of an EPI time series that should be thrown out, then motion-correct the remainder. We might then start to look at the effects of respiration and consider ways to further clean the time series.

There are also opportunities to upgrade the image acquisition. For a start we are using a two-channel (human) neck coil but the majority of the brain signal comes from the upper loop. This configuration negates the option to do simultaneous multislice (a.k.a. mutiband) EPI for fMRI. The size of the elements in the coil is suboptimal from a signal-to-noise (SNR) perspective, too. The open, circular geometry of the neck coil permits Greg to scan a wide range of dog sizes with the same equipment so I consider that a fixed specification for now. A coil mounted somehow to the dog's head has been considered and rejected because of the need for cabling. Cables would likely be too restrictive and/or get broken. An inductively-coupled head-mounted coil (which doesn't require a cable) might work, but it's a lot more complicated to design and getting multiple channels would be a problem. The option I've considered most seriously is a flexible "blanket" array coil, say 16 channels, that could be attached to a custom-built rigid former that allows the blanket to be mounted firmly as either a full or semicircle, easily mimicking the geometry of the current neck coil setup. A 16-channel array coil would permit a slice acceleration factor of at least R=2, perhaps even R=3 for SMS-EPI. It's something for me to consider as I see different RF coil designs become available.

I am also interested to know if we can acquire other types of functional scan from some of the better performing dogs. Both arterial spin labeling (ASL) and resting-state fMRI scans are highly motion-sensitive, but I'm fairly sure that some of the dogs would be able to remain as still as a human for the four or five minutes both of these types of scans require. Training in the mock scanner will tell us where the limits are. It might be appropriate to acquire ASL or rs-fMRI in shorter blocks with aggressive scrubbing of spoiled volumes. Scrubbing for rs-fMRI may change the nature of the data and I'm generally not a fan, unlike for task-based fMRI which can be considered as individual task blocks for statistical testing. But it could work for ASL. We shall have to try it and see.

At last month's Human Brain Mapping conference in Seattle, a poster by Harvard scientists Stephanie McMains and Ross Mair (poster 3412) showed yet more evidence that the benefits of a 32-channel coil for fMRI at 3 T aren't immediately obvious. Previous work by Kaza, Klose and Lotze in 2011 (doi: 10.1002/jmri.22614) had suggested that the benefits were regional, with cortical areas benefiting from the additional signal-to-noise ratio (SNR) whereas the standard 12-channel coil was superior for fMRI of deeper structures such as thalamus and cerebellum. The latest work by McMains and Mair confirms an earlier report from Li, Wang and Wang (ISMRM 17th Annual Meeting, 2009. Abstract #1614) that showed spatial resolution also affects the benefit, if any. In a nutshell, if a typical voxel resolution of 3 mm is used then the 32-channel coil provides no benefit over a 12-channel coil. The 32-channel coil was best only when resolution was pushed to 2 mm, thereby pushing the SNR down towards the thermal noise limit, or when using high acceleration, e.g. GRAPPA with acceleration, R > 2.

What's going on? In the first instance we need to think about the regimes that limit fMRI at different spatial resolutions. In the absence of subject motion and physiologic noise, the SNR of an EPI voxel will tend towards a thermal noise-limiting regime as it gets smaller. Let's assume a fairly typical SNR of 60 for a voxel that has dimensions 3.5x3.5x3.5 mm^3, as detected by a 12-channel head coil at 3 T. If we shrink the voxel to 3x3x3 mm^3 the SNR will decrease by ~27/43, to about 38, while if we shrink to 2x2x2 mm^3 the SNR will decrease to about 11. (Here I am assuming that all factors affecting N are invariant to resolution while S scales with voxel volume, which is sufficient for this discussion.) If we decrease the voxels to 1.5x1.5x1.5 mm^3 the SNR decreases to below five. The SNR is barely above one if we push all the way to 1x1x1 mm^3 resolution, which is why you don't often see fMRI resolution better than 2 mm at 3 T. Thus, if high spatial resolution is the goal then one needs to boost the SNR well beyond what we started of with to achieve a reasonable image. Hence the move to larger phased-array receive coils.

So that's the situation when the thermal noise is limiting. This is generally the case for anatomical MRI, but does it apply to fMRI? If something else is limiting - either physiologic noise or subject motion - then increasing the raw SNR may not help as expected. In fMRI we are generally less concerned with true (white) thermal noise than we are with erroneous modulation of our signal. It's not noise so much as it is signal changes of no interest. For this reason, Gonzalez-Castillo et al. (doi: 10.1016/j.neuroimage.2010.11.020) recently proposed using a very low flip angle in order to minimize physiologic noise while leaving functional signal changes unchanged.

From ISMRM e-poster 3352, available as a PDF via this link.

What if we can't even attain the physiologic noise-limiting regime? It's quite possible to be in a subject motion-limiting regime, as anyone who has run an fMRI experiment can attest. In that case, the use of a high dimensional array coil (of 32 channels, say) could actually impose a higher motion sensitivity on the time series than it would have had were it detected by a smaller array coil (of 12 channels, say), due to the greater receive field heterogeneity of the 32-channel coil. This was something a colleague and I considered last year, in an arXiv paper (http://arxiv.org/abs/1210.3633) and accompanying blog post. In an e-poster at this year's ISMRM Annual Meeting (abstract #3352; a PDF of the slides is available via this Dropbox link) we simulated the effects of motion on temporal SNR (tSNR), as well as the potential for spurious correlations in resting-state fMRI, when using a 32-channel coil. In doing these simulations we assumed perfect motion correction yet there were still drastic effects, as the above figure illustrates.

Whether the equivocal benefits of a 32-channel coil for routine fMRI (that is, using 3-ish mm voxels) are due to enhanced motion sensitivity, higher physiologic noise or some other factor I'm not in a position to say with any certainty. My colleagues and I, and others, are investigating ways that we might reduce the effects of receive field contrast on motion correction. The use of a prescan normalization is one idea that might help, at least a bit. The process has many assumptions and potential flaws, but it may offer the prospect of getting back some of what might be lost courtesy of the enhanced motion sensitivity. We simply don't know yet. The bigger problem, however, seems to be that a heterogeneous receive field contrast will impart motion sensitivity on a time series even if motion correction were perfect. Strong receive field heterogeneity, of the sort exhibited by a 32-channel head coil, is a killer if the subject moves.

Unless you are attempting to use highly accelerated parallel imaging (in particular the multiband sequences) and/or pushing your voxel size towards 2 mm, then, you're almost certainly better off sticking with the 12-channel coil as far as fMRI performance is concerned. Other scans, in particular anatomical scans and perhaps some diffusion-weighted scans, may benefit from larger array coils (because these scans may be in the thermal noise-limiting regime), but each application will need to be verified independently.

This is cool, publicly available test-retest pilot data sets using MB-EPI and conventional EPI on the same subjects courtesy of Nathan Kline Institute:



What's available:

• R-mfMRI (TR = 645msec; voxel size = 3mm isotropic; duration = 10 minutes)
• R-mfMRI (TR = 1400msec; voxel size = 2mm isotropic; duration = 10 minutes)
• R-fMRI (TR = 2500msec; voxel size = 3mm isotropic; duration = 5 minutes)
• Diffusion Tensor Imaging (137 direction; voxel size = 2mm isotropic)

The acquisition protocols are available as PDFs via the links given in the release website (and copied here). I like that they restricted the acceleration (MB) factor to four. I also like that the 3 mm isotropic MB-EPI data acquired at TR=645 ms used full Fourier acquisition (no partial Fourier) and an echo spacing of 0.51 ms. The former may help with signal in deep brain regions as well as frontal and temporal lobes, while the latter avoids mechanical resonances in the range 0.6-0.8 ms on a Trio, and also keeps the phase encode distortion reasonable.

There are already studies coming out that use these data sets, such as this one by Liao et al (which is how I learned of their existence). I don't yet know which reconstruction version was used for these data sets, but those of you who are tinkering should be aware that the latest version from CMRR, version R009a, has significantly lower artifacts and less smoothing than prior versions:

MB-EPI using CMRR sequence version R008 on a Siemens Trio with 32ch coil. MB=6, 72 slices, TE=38 ms, 2 mm isotropic voxels.

MB-EPI using CMRR sequence version R009a on a Siemens Trio with 32ch coil. MB=6, 72 slices, TE=38 ms, 2 mm isotropic voxels.

The bubbles visible in the bottom image of a gel phantom are real. The other intensity variations are artifacts. In both images one can easily make out the receive field heterogeneity of the 32-channel head coil.

----
Note added post publication

From Dan Lurie (@dantekgeek): We’re also collecting/sharing data from 1000 subjects using the same sequences, plus deep phenotyping http://fcon_1000.projects.nitrc.org/indi/enhanced/

PFUFA Part Fourteen introduced the idea of acquiring partial k-space and explained how the method, hereafter referred to as partial Fourier (pF), is typically used for EPI acquisitions. At this point it is useful to look at some example data and to begin to assess the options for using pF-EPI for experiments.


Image smoothing

The first consequence of using pF is image smoothing. It arises because we've acquired all of the low spatial frequency information twice - on both halves of k-space - but only half of some of the high spatial frequency information. We've then zero-filled that part of k-space that was omitted. This has the immediate effect of degrading the signal-to-noise ratio (SNR) for the high spatial frequencies that reside in the omitted portion of k-space. (PFUFA Part Eleven dealt with where different spatial frequencies are to be found in k-space.) Thus, the final image has less detail and is smoother than it would have been had we acquired the full k-space matrix, and because of the smoothing the final image SNR tends to be higher for pF-EPI than for the full k-space variant.

It was surprising to me that pF-EPI has higher SNR - due to smoothing - than full Fourier EPI in spite of the reduced data sampling in the acquisition. Conventional wisdom, which is technically correct, states that acquiring less data will degrade SNR. To understand this conundrum, we can think of pF as being like a square filter applied asymmetrically to the phase encoding dimension of an EPI obtained from a complete k-space acquisition. Indeed, as we start to evaluate the costs and benefits of pF for EPI we should probably be thinking about a minimum of a three-way comparison. Firstly, we obviously want to compare our pF-EPI to the full k-space alternative having the same nominal resolution. But we should also consider whether there is any advantage over a lower resolution EPI with full k-space coverage, too. Why? Because this lower resolution version is, in effect, what you get when partial Fourier is applied symmetrically, i.e. when the high spatial frequencies are omitted from both halves of the phase encoding dimension!

Let's do our first assessment of pF on a phantom. There are four images of interest: the full k-space image, two versions of pF - omitting the early or the late echoes from the echo train - and, for the sake of quantifying the amount of smoothing, a lower resolution full k-space image which is tantamount to omitting both the early and late echoes. (See Note 1.) From this point on I'm going to refer to omission of the early and late echo variants as pF(early)-EPI and pF(late)-EPI, respectively.

Images acquired from a structural phantom with a 12-channel head coil on a Siemens Trio. All parameters except the phase encode k-space sampling were fixed. Top left: 64x64 full Fourier EPI. Top right: 64x48 full Fourier EPI. Bottom left: 6/8ths pF(early)-EPI, reconstructed to 64x64. Bottom right: 6/8ths pF(late)-EPI, reconstructed to 64x64. (Click image to enlarge.)

I'm afraid it's not immediately clear, but if you look carefully you should be able to see that there are small features - a group of little dots in the middle of the central circle, for example - that are better resolved in the 64x64 full Fourier image than in either of the partial Fourier variants. The 64x48 matrix image is smoothest of all, as we would expect. It's also interesting to note that (Gibbs) ringing is prominent in the 64x64 matrix but much less so in the other three images. This prominence is another consequence of improved spatial resolution: ringing is always present to some extent because our pixels are, strictly speaking, sinc-shaped rather than square. Hard edges, as in this phantom, tend to exhibit the strongest ringing, and the higher the resolution the better the ringing is defined. (There's a review of ringing in this post.)


In fMRI we are interested in temporal stability, of course, so let's take a look at how partial Fourier affects temporal SNR (TSNR) when all other parameters (including TE) are held constant:

TSNR images derived from 100 volumes of EPI acquired from a structural phantom with a 12-channel head coil on a Siemens Trio. All parameters except the phase encode k-space sampling were fixed. Top left: 64x64 full Fourier EPI. Top right: 64x48 full Fourier EPI. Bottom left: 6/8ths pF(early)-EPI, reconstructed to 64x64. Bottom right: 6/8ths pF(late)-EPI, reconstructed to 64x64. (Click image to enlarge.)

The TSNR for the regions of interest in the above figure are as follows:

Top left            Full 64x64             TSNR = 343
Top right          Full 64x48             TSNR = 437
Bottom left      6/8ths pF(early)     TSNR = 435
Bottom right    6/8ths pF(late)       TSNR =  425

(Note: this is a throwaway comparison, for the purposes of illustration only! Please don't take the numbers you see here as absolutes, I am simply showing the effects of smoothing via SNR and TSNR because it may be difficult to see the smoothing on the limited details in these phantom images.)

The low resolution image (64x48) has higher TSNR than the higher resolution image (64x64 full k-space). We should expect a boost in SNR (and hence TSNR in a stationary object) from 343 to (64/48 x 343) = 457 because of the difference in voxel size. The observed TSNR of 437 isn't too far off.

Where things get slightly more interesting is for the pF-EPI variants. Conventional wisdom states that using partial Fourier will degrade the SNR in an image because less data is being recorded for an image that has the same nominal spatial resolution. For pF-EPI, however, the effect of smoothing (that is, the broadened point-spread function for the pixels) outweighs the signal-reducing effect of acquiring less data. Indeed, the observed TSNR is very close to that for the 64x48 full Fourier acquisition, indicating that the smoothing function is pronounced.

What about differences between the early and late echo variants of pF-EPI? Omitting the early echoes seems to boost TSNR very slightly more than omitting the late echoes, which is counter-intuitive because the early echoes will almost certainly have higher signal than the late echoes. Whether the difference in TSNR significant I won't get into because the difference is quite small and in other ROIs (not shown) the TSNR is almost identical. Besides, as you'll see below, there are other differences that might subjugate any smoothing differences. So, what's important at this juncture is that we recognize that the use of partial Fourier - omitting either the early or the late echoes - generates considerable image smoothing for EPI reconstructed with zero filling of the missing k-space.


Before we leave the smoothing issue, let's take a quick look at the effects on brain data since that's probably your interest. (My apologies, I didn't acquire the 64x48 full Fourier option from the brain. I'll do so for the next post, when I consider different pF-EPI schemes for fMRI.) Here's how 64x64 matrix full Fourier EPI compares to early and late 6/8ths pF-EPI variants:

Left: Full Fourier EPI acquired and processed as a 64x64 matrix. Center: 6/8ths pF(early)-EPI reconstructed to a 64x64 matrix with zero filling. Right: 6/8ths pF(late)-EPI reconstructed to a 64x64 matrix with zero filling. (Click picture to enlarge.)

If you have a good eye you may be able to see that the full Fourier acquisition, on the left, has finer detail than either of the pF-EPI options. I haven't quantified the SNR because it is highly region-dependent. (I cover the TSNR for these three acquisitions below.) On the basis of the phantom data above, however, we should expect the SNR to be increased for the pF variants entirely due to the smoothing effect. Whether this smoothing is acceptable or not for fMRI will be covered in the next post. Before we can make that determination we need to consider something else.


Signal dropout

Not all brain regions will see increased SNR because of smoothing. Some regions will see a degradation of SNR as a result of enhanced dropout. The origin of this effect was explained in PFUFA Part Fourteen. It is a consequence of the signals "falling off the edge" of the (curtailed) k-space plane because of magnetic susceptibility gradients.

Here are some brain images using 64x64 full Fourier EPI compared to 6/8ths pF(early)-EPI and 6/8ths pF(late)-EPI:

Left: 64x64 full Fourier EPI. Center: 6/8ths pF(early)-EPI. Right: 6/8ths pF(late)-EPI. (Click image to enlarge.)

Omitting the early echoes tends to enhance signal dropout in the temporal and frontal lobes (red and yellow arrows) while omitting the late echoes preserves temporal and frontal lobes but causes enhanced dropout in deep brain regions (blue arrows). This is interesting because it suggest that we have a degree of flexibility over where we pay the penalty for using partial Fourier EPI. I'll return to this issue later on, and in a subsequent post, because when we start to consider all the costs and benefits of pF-EPI we need to consider other parameters that might be changed in concert, such as the phase encoding direction, TE and the slice thickness.

Let's finish up this first look at pF-EPI in brain by assessing the TSNR. These images were obtained from 100 volumes with TE=22 ms and TR=2000 ms. All parameters were constant except the degree of partial Fourier sampling in the phase encoding dimension:

TSNR for different EPI sampling schemes. Left: 64x64 full Fourier EPI. Center: 6/8ths pF(early)-EPI. Right: 6/8ths pF(late)-EPI. (Click image to enlarge.)

In the figure I've included an ROI so that we can do a throwaway quantitative comparison. (As before, don't over-interpret what you see.)

Left            Full 64x64              TSNR =100
Center       6/8ths pF(early)      TSNR =115
Right          6/8ths pF(late)       TSNR = 123

All we can state with confidence is that the full Fourier images show a TSNR that is lower than either pF-EPI variants because of the smoothing, and that there are regions that have far lower - approaching zero - TSNR for the pF-EPI, due to the enhanced dropout that we saw above. Nothing new here, except one thing to note in passing: there doesn't appear to be a substantial difference in motion sensitivity when using pF-EPI. The smoothing-induced boost in TSNR is preserved in the brain images as it was in the stationary phantom images. This is as we expect because all we're doing is shortening a single-shot acquisition. (See Note 2.)


Options with partial Fourier

In the comparisons presented here I intentionally fixed all parameters except the partial Fourier scheme. That way you were able to get a sense of the direct costs or benefits of a partial Fourier scheme. But there are at least three other parameters that should be considered when setting up a partial Fourier scheme: (i) omitting the early echoes will permit a shorter minimum TE, (ii) omitting the late echoes will permit faster acquisition (i.e. more slices per TR) even when TE is unchanged, and (iii) the phase encoding direction makes "early" and "late" a relative property of the echo train. I'm going to leave consideration of these issues for the next post, when I will look at setting up pF-EPI for an fMRI experiment. None of these options is trivial. The TE affects BOLD sensitivity, the TR affects statistical power and brain coverage, while the phase encoding direction establishes whether distortions will be a stretch or a compression in a particular brain region. All of these issues interact depending on what parameters we change having selected a particular pF-EPI option, and the  optimal combination of parameters will depend on the brain region(s) of interest in your experiment.




Notes:

1.  Please note that Siemens' product EPI sequences don't have a way to select the late echoes for their partial Fourier options. The early echoes are always omitted with these sequences. However, in the coming months/years I hope to offer to research users a modified sequence that has early/late echo omission as an option, along with a host of other small tweaks that can be useful for fMRI. As soon as this sequence is available for distribution I'll be sure to blog about it.

2.  Pedants will note that the actual motion sensitivity is a function of the underlying image contrast so that, in strict terms, the pF-EPI and full Fourier EPI scans do have different motion sensitivities. But this difference also exists for different brain shapes, different head orientations, different RF flip angles (because flip angle and TR establish the T1-based image contrast), etc. Might the motion sensitivity actually be reduced with pF-EPI,? It seems unlikely. Although the per slice time is decreased with pF-EPI, we also have to recognize that the effects of magnetic susceptibility are changed, too. So, what we gain with speed on the one hand we might give up with susceptibility contrast effects - signal dropout in other words - on the other. I really couldn't say whether these effects will be offsetting or not, and as far as I know nobody has ever assessed it. My bet would be that proving a systematic difference would be difficult because I suspect the motion sensitivity differences would be tiny.

Have you ever wondered why your fMRI scanner is the way it is? Why, for example, is the magnet typically operated at 1.5 or 3 T, and why is there a body-sized transmission coil for the RF? The prosaic answer to these questions is the same: it's what's for sale. We are fortunate that MRI is a cardinal method for radiology, and this clinical utility means that large medical device companies have invested hundreds of millions of dollars (and other currencies) into its development. The hardware and pulse sequences required to do fMRI research aren't fundamentally different from those required to do radiological MRI so we get to use a medical device as a scientific instrument with relative ease.

But what would our fMRI scanners look like today had they been developed as dedicated scientific instruments, with little or no application to something as lucrative as radiology? Surely the scanner-as-research-device would differ in some major ways from that which is equally at home in the hospital or the laboratory. Or would it? While it's clear that the fMRI revolution of the past twenty years has ridden piggyback on the growing clinical importance of diffusion and other advanced anatomical imaging techniques, what's less obvious is the impact of these external factors on how we conduct functional neuroimaging today. State-of-the-art fMRI might have looked quite different had we been forced to develop scanners explicitly for neuroscience.


"I wouldn't start from here, mate."

This week's interim report from the BRAIN Initiative's working group is an opportunity for all of us involved in fMRI to think seriously about our tools. We've come a long way with BOLD contrast to be sure, even though we don't fully understand its origins or its complexities. Should I be delighted or frustrated at my capacity to operate a push-button clinical machine at 3 T in order to get this stuff to work? It's undoubtedly convenient, but at what cost to science?

I can't help but wonder what my fMRI scanner might look like if it was designed specifically for task. Would the polarizing magnet be horizontal or would a subject sit on a chair in a vertical bore? How large would the polarizing magnet be, and what would be its field strength? The gradient set specifications? And finally, if I'm not totally sold on BOLD contrast as my reporting mechanism for neural activity, what sort of signal do I really want? In all cases I am especially interested in why I should prefer one particular answer over the other alternatives.

Note that I'm not suggesting we all dream of voltage-sensitive contrast agents. That's the point of the BRAIN Initiative according to my reading of it. All I'm suggesting is that we spend a few moments considering what we are currently doing, and whether there might be a better way. Unless there has been a remarkable set of coincidences over the last two decades, the chances are good that an fMRI scanner designed specifically for science would have differed in some major ways from the refined medical device that presently occupies my basement lab. There would be more duct tape for a start.

I've been dabbling in some ultralow field (ULF) MRI over the past several years, trying first to get functional brain imaging to work (more on that another day, perhaps) and more recently looking at the contrast properties of normal and diseased brains. We detect MR signals at less than three times the earth's magnetic field (of approximately 50 microtesla) using an ultra-sensitive superconducting quantum interference device (SQUID). The system is usually referred to as "The Cube" on account of the large aluminum box surrounding the entire apparatus; it provides magnetic shielding for the SQUID. But my own nickname for the system is CALAMARI - the CAL Apparatus for MAgnetic Resonance Imaging. Deep-fried rings or grilled strips, it's all good. Anyway, should you wish to know more about this home-built system and what it might be able to do, there's a new paper (John Clarke's inaugural article after being elected to the NAS) now out in PNAS. At some point I'll put up more blog posts on both anatomical and functional ULFMRI, and go over some of the work that's being done at high fields (1.5+ T) that may be relevant to ULFMRI.

Back in August I did a post on the experimental consequences of using partial Fourier for EPI. (An earlier post, PFUFA Part Fourteen introduces partial Fourier EPI.) The main point of that post was to demonstrate how, with all other parameters fixed, there are two principal effects on an EPI obtained with partial Fourier (pF) compared to using full phase encoding: global image smoothing, and regionally enhanced signal dropout. (See Note 1.)

In this post I want to look a little more closely at how pF-EPI works in practice, on a brain, with fMRI as the intended application, and to consider what other parameter options we have once we select pF over full k-space. I'll do two sets of comparisons. In the first comparison all parameters except the phase encoding k-space fraction will be fixed so that we can again consider the first stage consequences of using pF. In the second comparison each pF-EPI scheme will be optimized in a "maximum performance" test. The former is an apples to apples comparison, with essentially one variable changing at a time, whereas the latter is how you would ordinarily want to consider the pF options available to you.


Why might we want to consider partial Fourier EPI for fMRI anyway?

If we assume a typical in-plane matrix of 64 x 64 pixels, an echo spacing (the time for each phase-encoded gradient echo in the train, as explained in PFUFA Part Twelve) of 0.5 ms and a TE of 30 ms for BOLD contrast then it takes approximately 61 ms to acquire each EPI slice. (See Note 2 for the details.) The immediate consequence should be obvious: at 61 ms per slice we will be limited to 32 slices in a TR of 2000 ms. If the slice thickness is 3 mm then the total brain coverage in the slice dimension will be ~106 mm, assuming a 10% nominal inter-slice gap (i.e. 32 x 3.3 mm slices). With axial slices we aren't going to be able to cover the entire adult brain. We will have to omit either the top of parietal lobes or the bottom of the temporal lobes, midbrain, OFC and cerebellum. Judicious tilting might be able to capture all of the regions of primary interest to you, but we either need to reduce the time taken per slice or increase the TR to cover the entire brain.

Partial Fourier is one way to reduce the time spent acquiring each EPI slice. There are two basic ways to approach it: eliminate either the early echoes or the late echoes in the echo train, as described at the end of PFUFA: Part Fourteen. Eliminating the early echoes doesn't, by itself, save any time at all. Only if the TE is reduced in concert is there any time saving. But omitting the late echoes will mean that we complete the data acquisition for the current slice earlier than we would for full Fourier sampling, hence there is some intrinsic speed benefit. I'll come back to the time savings and their consequences later on. Let's first look at what happens when we enable partial Fourier without changing anything else.


Image quality assessment for pF-EPI

Our gold standard will be full k-space EPI with a 64 x 64 matrix. For this post I am only going to use the 6/8ths partial Fourier option, meaning that one quarter (2/8ths) of the phase encoding k-space will be omitted from the acquisition. Thus, we will have acquired 48 of 64 phase encode lines and will simply zero fill the missing lines prior to 2D FT of a (synthetic) 64 x 64 matrix. Again, see PFUFA: Part Fourteen for an introduction to partial Fourier EPI if this vernacular leaves you cold.

As we saw previously, one effect of acquiring a partial k-space is image smoothing. Which immediately begs the question: why bother using pF at all, and why not just reduce the matrix size (symmetrically) instead? So, one comparison we want to make, specifically to evaluate image smoothing, is the acquisition of a full Fourier 64 x 48 matrix lower resolution EPI. In this case we acquire k-space symmetrically in the phase encoding dimension; we're leaving off 1/8th of the early and 1/8th of the late echoes compared to the full 64 x 64 matrix acquisition.

As we've seen previously, there are two options for 6/8ths pF-EPI. We can omit the early or the late phase encoded echoes, as illustrated in this figure (see Note 3):



I shall try always to refer consistently to the former as pF(early) and the latter as pF(late), but in some of the images you may notice that in practice I tend to refer to the former as simply pF while the latter is pFrev, for "reversed" pF. So if you see "rev" or "reversed" in any data just think "late" instead.

I also want to emphasize here that early and late (or reversed) are designations made relative to the phase encoding direction that's being used. For axial slices the Siemens default is to use anterior-posterior (A-P) phase encoding. (I've noted previously that GE uses P-A by default.) If the imaging gradients were perfect and there were no magnetic susceptibility gradients across the head then omitting the late echoes for A-P phase encoding would be tantamount to omitting the early echoes for P-A phase encoding. But we don't have a perfect system and we shall therefore want to do a separate set of comparisons for P-A phase encoding, distinct from those for A-P. The imperfections? Mostly, it's those pesky magnetic susceptibility gradients that cause distortion and dropout. The phase encoding dimension dictates the direction of distortion and you will almost certainly have a preference. Also, the local regions that exhibit enhanced signal dropout will differ with phase encoding direction.

Disclaimer:  Do not, under any circumstances, treat these results as a validation of either of the pF variants!!! All I offer is a starting point for you to ponder your alternatives. Unless and until someone provides a validation of pF you should remain skeptical. At a minimum, you would want to conduct a thorough pilot experiment before selecting a pF variant for a full-blown fMRI experiment.


Disclaimer over, here is our first set of comparisons, in this case using A-P phase encoding:

EPI with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

All parameters except the phase encoding fraction are constant: Siemens TIM/Trio, 12-channel head coil, TR = 2000 ms, TE = 22 ms, FOV = 224 mm x 224 mm, slice thickness = 3 mm, inter-slice gap = 0.3 mm, echo spacing = 0.5 ms, bandwidth = 2232 Hz/pixel, flip angle = 70 deg. Each EPI was reconstructed as a 64x64 matrix however much actual k-space was acquired, and any omitted portions were zero-filled prior to 2D FT.

Let's zoom in a bit to get a better look at those slices that typically exhibit regions of dropout:

EPI with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Additional dropout is evident for both of the pF options as well as for the low resolution full k-space EPI, compared to the 64 x 64 reference images. Temporal lobes and midbrain are affected most, consistent with the brain data shown in the last post on pF-EPI. (See Note 4 for more information on the effects of resolution on dropout.)

What about image smoothing? It's hard to see on brains, but there are a couple of slices on which you can, if you have a good eye, just about discern the different edge detail:

EPI with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

We aren't exclusively interested in the appearance of EPI or the brightness of a particular region when doing fMRI, however. We are using time series acquisitions so we need to consider motion sensitivity and the signal stability over time. So let's shift to assessing temporal images.

We can make a reasonable assessment of any differential motion sensitivity by looking at standard deviation images. Here are the results for the A-P phase encoding data from above, for fifty-volume time series acquisitions (100 secs of data) in each case:

Standard deviation images for fifty EPI time series acquisitions, with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

As we would expect for single shot EPI acquisitions, the motion sensitivity is fairly consistent. Using partial Fourier doesn't change the fact that the acquisition is still a single echo train acquired after a single slice-selective RF pulse. Thus, no one scheme exhibits vastly different variance than the others. There will probably be localized differences, however. In the case of partial Fourier, signals might be on the very edge of the k-space "cliff" and they might be one or other side of that drop with quite small subject movement. But determining relative performance requires regions-of-interest - something that will vary depending on your application - or some way to collapse the signal stability for the whole brain down to a single value, a process that might easily obscure subtle effects that are actually important. So, let's just accept that the motion sensitivity is operationally similar, and move on.

Of even more relevance to fMRI is the temporal SNR (tSNR), a handy proxy for signal level as well as stability. Here are voxelwise tSNR maps of the same fifty-volume time series as used in the standard deviation images above:

Temporal SNR images for fifty EPI time series acquisitions, with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Now we can see the full effects of image smoothing. The tSNR is higher for 64x48 full Fourier and both partial Fourier options compared to the 64x64 full Fourier baseline image. But we have netted "extra" SNR purely by smoothing the image, an effect we could get in post processing with a smoothing function applied to the 64x64 full Fourier image! Is there a difference between the 64x48 full Fourier and either of the pF options? In regions with good signal, not really. But where the sampling scheme enhances signal dropout compared to 64x64 full Fourier then we see the same holes in the tSNR images as we saw in the raw EPI above. What's gone is gone.

That about wraps up the quick and dirty visual assessment of the pF-EPI options using A-P phase encoding. There is another entire four-way comparison on offer, however: the same four sampling schemes but applied with P-A phase encoding! I've put all the data for P-A phase encoding into Appendix 1. Here, let's stick with A-P phase encoding but turn our attention to some comparisons when the timing parameters aren't held constant, which is far more realistic.


How fast can we make pF-EPI go?

Let's assume we have valid reasons for not wanting to increase TR beyond 2000 ms, and let's further assume that the gradients are being driven as fast as they can go. (In all of the data shown in this post the echo spacing is fixed at 0.5 ms.) We need a way to save some time if we are to acquire more slices in the specified TR. Partial Fourier is one option for saving time per slice.

By setting the parameters for "maximum performance" - meaning the use of minimum TE and as many slices as we can fit in TR=2000 ms - it turns out that we get 43 slices for the two pF options as well as for the 64x48 low-res option, compared to 37 slices for the 64x64 full k-space standard. But in achieving the 43 slices, only 6/8pF(late) uses the same TE=22 ms as the 64x64 standard. For pF(early) the TE is reduced to the minimum value of 14 ms while for 64x48 low-res the TE is reduced to 18 ms. Using longer than the minimum TE in either case results in fewer than 43 slices in TR=2000 ms.

For space considerations, and to allow you to make better comparisons on your own screens, I've put large matrix figures for each of the four options (all using A-P phase encoding) in Appendix 2. I'll note in passing that these images show the same effects of smoothing and regional dropout as the constant parameter comparisons above, and move on.

With the "maximum performance" parameter settings the motion sensitivity remains quite similar to the prior comparisons. This is as we should expect for single shot EPI; minor differences in TE won't have a large effect. Thus, the standard deviation images have comparable artifact levels for edges (due to motion), physiologic fluctuations and for N/2 ghosts:

Standard deviation images for "maximum performance" acquisitions with phase encoding set A-P.  Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

What about tSNR over a fifty volume time series?

Temporal SNR images for "maximum performance" acquisitions with phase encoding set A-P.  Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

As with the fixed parameter comparison, the effects of smoothing - enhanced tSNR - on the 64x48 images is especially noticeable compared to the 64x64 case. The tSNR is higher still for the 6/8pF(early) case, a combination of image smoothing plus the use of a shorter TE of 14 ms. The tSNR for 6/8pF(late) is comparable to that for 64x64 full Fourier, however. But this is only part of the story, which is why I'm avoiding quantitative comparisons for any one region. Let's consider dropout. We can see that dropout of midbrain signal is higher for 6/8pF(late) than it is for any of the other three options. Yet signal in temporal lobes is well preserved for 6/8pF(late), comparable to that in the full Fourier 64x64 images. Signal for temporal lobes in both the 64x48 full and 6/8pF(early) images show more dropout. Thus, if you were interested in auditory fMRI you might want to consider 64x64 full or 6/8pF(late), but if you're doing, say, hypothalamus then either 64x48 full or 6/8pF(early) look to be better candidates.

A comparison for "maximum performance" parameters but with P-A phase encoding is given in the appendices. The individual image mosaics are in Appendix 3 while the standard deviation and tSNR images for fifty-volume time series are in Appendix 4.


Summary

In the previous post on partial Fourier EPI you saw how partial Fourier affects a single image. In this post, the analysis was expanded to consider the effects on a time series and also different parameter combinations for time series acquisitions. What are the broader lessons to take away so far?

• Partial Fourier leads to image smoothing. It's important to note that any apparent gain in SNR (for otherwise fixed parameters) is due to the smoothing.
• Partial Fourier usually contributes to enhanced signal dropout, especially in the "problem" brain regions of midbrain, frontal lobes and temporal lobes where magnetic susceptibility gradients are worst. You may be able to select which regions exhibit worse dropout by judicious combination of phase encode direction and early or late echo omission.
• Omitting the early echoes from the gradient echo train can benefit EPI by permitting a shorter TE. If we use pF(early) and we don't shorten the TE then all we're really doing is giving up SNR, especially for the regions mentioned in the previous point. (Remember that there is still considerable BOLD (T2*) weighting during the EPI echo train. BOLD contrast isn't entirely dependent on the TE!)
• Omitting the late echoes from the gradient echo train doesn't change the minimum TE but it does permit faster acquisition of slices, i.e. more slices in TR.
• Partial Fourier doesn't make EPI more motion-sensitive. Strictly speaking, the image contrast does impact motion sensitivity and motion correction a little bit, but these factors are affected by many other parameters, too, such as the excitation flip angle.

To finish up, here is a little general guidance when considering partial Fourier EPI:

• Consider what you want to do with TE whenever you are assessing partial Fourier as an option.
• If you omit early echoes then you'll almost certainly want to reduce TE as well.
• If reduced dropout is your focus then you may want a reduced TE for its own sake, and perhaps thinner slices (see two points down).
• If you omit late echoes then the assumption is that you're aiming for more slices in TR.
• Even if you are happy with your slice coverage sans pF, using pF may permit a greater number of thinner slices for the same total coverage in the slice dimension. But you would have to determine whether there is net benefit from thinner slices versus the enhanced regional dropout mentioned already.
• Regarding regional dropout, you may have a degree of choice as to which signals are sacrificed in setting early or late echo omission by setting the phase encode gradient polarity, e.g. P-A instead of A-P. But there is a concomitant effect on distortion direction, too.

The next post in this series will consider partial Fourier EPI compared to alternative "go faster" options, in particular the use of GRAPPA. And then we'll shift focus to simultaneous multislice (SMS), aka multiband (MB) EPI.





Notes:

1.  In textbooks you will usually encounter a description of partial Fourier phase encoding that involves the decrease of image SNR because of the reduced signal averaging compared to a fully sampled k-space plane. Strictly speaking, it will be accurate. In practice, however, a loss of SNR with pF doesn't manifest in EPI the way the textbooks describe it. Instead, we tend to fnd an apparent increase of image SNR across most of the EPI, arising from the smoothing imposed by the 'zero filling' filter effect. Thus, a higher apparent SNR resulting from pF isn't a "real" SNR gain but comes from smoothing. You could get the same - or better - SNR from taking a fully sampled EPI and applying a smoothing function in post-processing. We do see decreased SNR but it tends to be regional. Signals from some brain areas 'fall off' the sampled k-space plane due to magnetic susceptibility gradients. Keep these points in mind when comparing the different SNR levels observed in the comparisons that follow.

2.  A real EPI pulse sequence was considered in PFUFA: Part Thirteen. In addition to sampling of the k-space plane with the repeated gradient echoes, there are also temporal overheads for each slice: fat suppression, slice selection, and a short crusher gradient at the end of each slice that eliminates any residual signal prior to the next slice (hopefully). For simplicity, let's assume that it takes a total of 15 ms to do a fat suppression pulse, the first half of a slice selection (the second half being accounted for within TE), and a short crusher gradient after each slice is acquired. This is the temporal overhead per slice. Next we need to determine the time taken to sample the 2D k-space plane.
         For a 64 x 64 matrix EPI with 0.5 ms echo spacing it takes 32 x 0.5 ms = 16 ms to reach the center of k-space, then a further 16 ms to reach the end of the in-plane information. The TE defines the center of k-space, however, so the mid-point of the 64 echoes has to be "parked" at TE. Thus, the first 32 echoes, taking 16 ms, can be acquired within the 30 ms allowed for TE. The latter 32 echoes take a further 16 ms after TE to acquire. Thus, the total time per slice is  TE + 16 ms + 15 ms (overhead) = 61 ms to acquire a single 2D plane. There may be small variations but this is a pretty good estimate.

3.  Siemens users, I'm afraid that you can only neglect the early echoes in the product EPI sequences such as ep2d_bold and ep2d_pace. I'm working on getting an early/late option into a subsequent product sequence, and/or making available a research version of ep2d_bold. Big, bureaucratic subject for another day. Right now the question is whether there's any benefit to having the early/late option at all!

4.  There is a general principle at work here: higher resolution for EPI - whether in-plane or thinner slices or both - will tend to reduce the extent of magnetic susceptibility gradients across a voxel and thus tend to reduce the dephasing causing signal loss. It's the same principle that was demonstrated for the slice thickness in the "Signal dropout" section of PFUFA: Part Twelve, but we can extend it to 3D. Now, there's no free lunch. In exchange for reducing the dephasing across a (smaller) voxel we lose the SNR on a volumetric basis; voxels with 2 mm sides produce base SNR that is less than one third that of voxels with 3 mm sides. And because we have smaller voxels we now have a potential brain coverage issue, especially in the slice dimension. Still, aiming for smaller voxels is one of the tactics for reducing dropout in EPI.


Appendix 1:

All parameters except the phase encoding fraction are constant: Siemens TIM/Trio, 12-channel head coil, TR = 2000 ms, TE = 22 ms, FOV = 224 mm x 224 mm, slice thickness = 3 mm, inter-slice gap = 0.3 mm, echo spacing = 0.5 ms, bandwidth = 2232 Hz/pixel, flip angle = 70 deg, phase encoding direction = P-A. Each EPI was reconstructed as a 64x64 matrix however much actual k-space was acquired:

EPI with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Again, we can zoom in to assess likely problem regions:

EPI with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

And check on smoothing:

EPI with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Standard deviation images for fifty-volume time series acquisitions with P-A phase encoding:

Standard deviation images for fifty EPI time series acquisitions, with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Temporal SNR images for fifty-volume time series acquisitions with P-A phase encoding:

Temporal SNR images for fifty EPI time series acquisitions, with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Appendix 2:

High resolution versions of the four "maximum performance" acquisitions with A-P phase encoding:

64x64 full Fourier, A-P phase encoding, 37 slices in TR=2000 ms, TE = 22 ms.

64x48 full Fourier, A-P phase encoding, 43 slices in TR=2000 ms, TE = 18 ms.

6/8pF(early), A-P phase encoding, 43 slices in TR=2000 ms, TE = 14 ms.

6/8pF(late), A-P phase encoding, 43 slices in TR=2000 ms, TE = 22 ms.

Appendix 3:

High resolution versions of the four "maximum performance" acquisitions with P-A phase encoding:

64x64 full Fourier, P-A phase encoding, 37 slices in TR=2000 ms, TE = 22 ms.

64x48 full Fourier, P-A phase encoding, 43 slices in TR=2000 ms, TE = 18 ms.

6/8pF(early), P-A phase encoding, 43 slices in TR=2000 ms, TE = 14 ms.

6/8pF(late), P-A phase encoding, 43 slices in TR=2000 ms, TE = 22 ms.

Appendix 4:

Standard deviation images for fifty-volume time series acquisitions with P-A phase encoding and "maximum performance" parameters:

Standard deviation images for "maximum performance" acquisitions with phase encoding set P-A.  Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

 Temporal SNR images for fifty-volume time series acquisitions with P-A phase encoding and "maximum performance" parameters:

Temporal SNR images for "maximum performance" acquisitions with phase encoding set P-A.  Top left: 64x64 full Fourier. Top right: 64x48 full Fourier. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

This is the final post in a short series concerning partial Fourier EPI for fMRI. The previous post showed how partial Fourier phase encoding can accelerate the slice acquisition rate for EPI. It is possible, in principle, to omit as much as half the phase encode data, but for practical reasons the omission is generally limited to around 25% before image artifacts - mainly enhanced regional dropout - make the speed gain too costly for fMRI use. Omitting 25% of the phase encode sampling allows a slice rate acceleration of up to about 20%, depending on whether the early or the late echoes are omitted and whether other timing parameters, most notably the TE, are changed in concert.

But what other options do you have for gaining approximately 20% more slices in a fixed TR? A common tactic for reducing the amount of phase-encoded data is to use an in-plane parallel imaging method such as SENSE or GRAPPA. Now, I've written previously about the motion sensitivity of parallel imaging methods for EPI, in particular the motion sensitivity of GRAPPA-EPI, which is the preferred parallel imaging method on a Siemens scanner. (See posts here, here and here.) In short, the requirement to obtain a basis set of spatial information - that is, a map of the receive coil sensitivities for SENSE and a set of so-called auto-calibration scan (ACS) data for GRAPPA - means that any motion that occurs between the basis set and the current volume of (accelerated) EPI data is likely to cause some degree of mismatch that will result in artifacts. Precisely how and where the artifacts will appear, their intensity, etc. will depend on the type of motion that occurs, whether the subject's head returns to the initial location, and so on. Still, it behooves us to check whether parallel imaging might be a better option for accelerating slice coverage than partial Fourier.


Deciding what to compare

Disclaimer: As always with these throwaway comparisons, use what you see here as a starting point for thinking about your options and perhaps determining your own set of pilot experiments. It is not the final word on either partial Fourier or GRAPPA! It is just one worked example.

Okay, so what should we look at? In selecting 6/8ths partial Fourier it appears that we can get about 15-20% more slices for a fixed TR. It turns out that this gain is comparable to using GRAPPA with R=2 acceleration with the same TE. To keep things manageable - a five-way comparison is a sod to illustrate - I am going to drop the low-resolution 64x48 full Fourier EPI that featured in the last post in favor of the R=2 GRAPPA-EPI that we're now interested in. For the sake of this comparison I'm assuming that we have decided to go with either pF-EPI or GRAPPA, but you should note that the 64x48 full Fourier EPI remains an option for you in practice. (Download all the data here to perform for your own comparisons!)

I will retain the original 64x64 full Fourier EPI as our "gold standard" for image quality as well as the two pF-EPI variants, yielding a new four-way comparison: 64x64 full Fourier EPI, 6/8pF(early), 6/8pF(late), and GRAPPA with R=2. Partial Fourier nomenclature is as used previously. All parameters except the specific phase encode sampling schemes were held constant. Data was collected on a Siemens TIM/Trio with 12-channel head coil, TR = 2000 ms, TE = 22 ms, FOV = 224 mm x 224 mm, slice thickness = 3 mm, inter-slice gap = 0.3 mm, echo spacing = 0.5 ms, bandwidth = 2232 Hz/pixel, flip angle = 70 deg. Each EPI was reconstructed as a 64x64 matrix however much actual k-space was acquired. Partial Fourier schemes used zero filling prior to 2D FT. GRAPPA reconstruction was performed on the scanner with the default vendor reconstruction program. (Siemens users, see Note 1.)


Image quality assessment

In this comparison the phase encoding direction is anterior-posterior (A-P), the Siemens default. (See Appendix 1, below, for a similar four-way comparison using P-A phase encoding.) There are 37 slices in TR=2000 ms, which is the maximum number of slices permitted by the full Fourier 64x64 matrix EPI. Here are the images after zooming to crop the uppermost two slices from each data set:

EPI with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

And here are the same images but zoomed so that we can get a better look at likely problem areas:

EPI with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Comparing the pF schemes to the full Fourier EPI first, we see the now familiar regions of enhanced dropout - primarily temporal lobes (and eyes!) for 6/8pF(early), midbrain for 6/8pF(late) - and also the smoother images arising from zero filling the partial Fourier EPIs.

The most immediate difference between the GRAPPA-EPI and the other three data sets is the reduced distortion in the A-P direction. Partial Fourier doesn't alter the amount of distortion whereas GRAPPA reduces distortion by the acceleration factor, R=2 in this case. The distortion is worst where the magnetic susceptibility gradients are worst, so the reduced distortion is most evident in the temporal lobes. Distortion of the frontal lobe signal is also halved but the benefit is less obvious because it appears that there might be additional dropout with the GRAPPA acquisition. Why the dropout should get worse isn't immediately obvious, but we can speculate that it's a reconstruction error arising from a mismatch between the ACS and this undersampled volume. Not a good sign.

It's time to look at the performance of GRAPPA in a time series. Here are the standard deviation images for 50-volume time series:

Standard deviation images for fifty EPI time series acquisitions, with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Uh-oh. Clearly, the temporal stability of the GRAPPA data is worse than all the other three schemes. The experienced subject was careful not to move during the ACS - for instance, he swallowed immediately before the start of the scan - and did his best not to move during the time series, too. Yet the frontal lobes in particular exhibit large standard deviations, and there is a pronounced ring around the circumference of the head for all slices. What does this do to the temporal SNR? Let's look:

Temporal SNR images for fifty EPI time series acquisitions, with phase encoding set A-P and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

As we might expect when the standard deviation is high, the tSNR for the GRAPPA scheme is reduced below that for the two partial Fourier schemes as well as the reference 64x64 full Fourier EPI. The price for the speed gain seems to be about half of the tSNR, according to the region-of-interest selected in this throwaway comparison.

It is important to note that the performance of GRAPPA is quite variable. In Appendix 1, below, you will find the same four-way comparison but with the phase encoding direction reversed, to P-A. In those images you'll see that the GRAPPA stability is still the worst of the four, but it isn't quite as bad as in the A-P data above. This is the problem when using GRAPPA: just one head movement - a swallow, say - can have very severe consequences for the overall time series. More on the costs and benefits below.


Going at maximum speed

For the purposes of this post, the motivation for adopting partial Fourier or GRAPPA is to attain more slices in the TR. So let's look at the time series statistics when the TE is reduced as far as possible in order to permit the maximum number of slices in TR = 2000 ms. (Reducing TE to the minimum attainable isn't always what you would want to do for BOLD contrast, but I'm doing it here to get the maximum number of slices.) Except for the TE and the number of slices, all other parameters were left set at the values given previously.

The good news is that GRAPPA with R=2 acceleration and a minimum TE of 14 ms allows a whopping 52 slices in TR = 2000 ms! Mission accomplished, right? Perhaps. If you don't mind giving up that temporal stability. Here is the four-way comparison of standard deviations for fifty-volume time series acquisitions:

Standard deviation images for "maximum performance" acquisitions with phase encoding set A-P. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

It looks like we have a similar motion sensitivity as before. The shorter TE for GRAPPA will enhance the image SNR and this should translate into improved temporal SNR in the absence of motion. We end up seeing a net loss, however, because of the motion sensitivity:

Temporal SNR images for "maximum performance" acquisitions with phase encoding set A-P. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Note that the region-of-interest in the top-right matrix doesn't precisely match the other three because the high number of slices caused the image display to shift in Osirix. I did my best to get a similar region. Still, it is clear that there is a rather large global penalty in the GRAPPA data compared to the partial Fourier options.

Appendix 2 contains the same four-way comparison but with the phase encoding reversed, to P-A. GRAPPA again performs the worst of the bunch.


Lessons learned

It appears that using GRAPPA with R=2 is quite costly in terms of reduced temporal SNR. In the "maximum performance" test I reduced the TE to the minimum of 14 ms, a situation that probably isn't something that you would do for fMRI. You might reduce the TE to around 20 ms for BOLD.

At a TE of around 20 ms the major apparent benefit of GRAPPA - more slices per TR than for partial Fourier - becomes marginal, yet it comes at the cost of greatly enhanced motion sensitivity. To me, it doesn't seem worth the cost for such a relatively small gain in imaging speed. If the objective is to tease out an additional 20% more slices in TR then it appears that partial Fourier EPI is the better (safer) alternative.

So, what about going even faster? Why stop at R=2 for GRAPPA? It is certainly possible to use R=3 or 4 with large phased-array coils, but at the cost of further enhancement of motion sensitivity. What's more, using in-plane acceleration gets us percentages of speed increase but it doesn't get us factors. What if you wanted to get twice as many slices in a fixed TR, or three times as many? In that case you should probably focus on the slice dimension itself and accelerate it directly, using simultaneous multi-slice (aka multi-band) EPI. That will be the subject of the next post.




Notes:

1.  On VB17A software, and previously on VB15, the product EPI sequence uses a single ACS for R=2 accelerated GRAPPA EPI. This means that there is a mismatch between the k-space step size for the ACS and the step size - twice as big - for the undersampled EPI of the time series. Such a mismatch leads to reconstruction errors whenever there are appreciable magnetic susceptibility gradients acting to distort the phase encoding. On my scanner we therefore use a tweaked version of ep2d_bold for which the correct R-shot ACS is acquired for R=2. Note, however, that Siemens does correctly acquire 3-shot and 4-shot ACS for R=3 and 4. It's just R=2 that has the potential mismatch. See the introduction section of this arXiv paper for more details.



Appendix 1:

All parameters except the phase encoding fraction are constant: Siemens TIM/Trio, 12-channel head coil, TR = 2000 ms, TE = 22 ms, FOV = 224 mm x 224 mm, slice thickness = 3 mm, inter-slice gap = 0.3 mm, echo spacing = 0.5 ms, bandwidth = 2232 Hz/pixel, flip angle = 70 deg, phase encoding direction = P-A. Each EPI was reconstructed as a 64x64 matrix however much actual k-space was acquired:

EPI with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Zoomed to show likely problem regions:

EPI with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Standard deviation images for fifty-volume time series acquisitions with P-A phase encoding:

Standard deviation images for fifty EPI time series acquisitions, with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Temporal SNR images for fifty-volume time series acquisitions with P-A phase encoding:

Temporal SNR images for fifty EPI time series acquisitions, with phase encoding set P-A and all parameters held constant except for the phase encode sampling scheme. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Appendix 2:

Standard deviation images for fifty-volume time series acquisitions with P-A phase encoding and "maximum performance" parameters:

Standard deviation images for "maximum performance" acquisitions with phase encoding set P-A. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Temporal SNR images for fifty-volume time series acquisitions with P-A phase encoding and "maximum performance" parameters:

Temporal SNR images for "maximum performance" acquisitions with phase encoding set P-A. Top left: 64x64 full Fourier. Top right: GRAPPA-EPI with R=2 acceleration. Bottom left: 6/8pF(early). Bottom right: 6/8pF(late). (Click image to enlarge.)

Those of you who follow me on Twitter may have noticed that I've been scanning some post mortem brains of Cetacea over the past year or so. That's whales, dolphins and porpoises to you and me. The brains come in all shapes and sizes, from a rather tiny Amazon river dolphin, about the size of a fist, to fin and sei whale brains that are so wide they have to be inserted sideways, hemisphere-first, into the 3 T (human) head coil. The conditions and ages of the brains vary tremendously as well. Some have been fixed in formaldehyde for decades yet yield remarkably decent signal, others have been stored in ethanol and are as hard as rubber with T2 to match. A few months ago we obtained a recently deceased fresh brain of a white-sided dolphin which we were able to scan within about twelve hours of its demise. The image quality was magnificent.

What do we plan to do with all the post mortem data? That is still being formulated. Initial motivation for the project came from some Berkeley anthropologists with an interest in comparative neuroanatomy across higher mammalian species. Coincidentally, Greg Berns' group at Emory has recently produced a nice example of dolphin brain tractography and his recent study is a good example of what might be done in future. There's a commentary on Greg's study here, and an example image from his paper below. We are now determining how we might combine resources, share data and all that good stuff. More on what will be available to whom and when as we progress.

From: Berns et al. DOI 0.1098/rspb.2015.1203

In any case, after I posted the white-sided dolphin MRI to Twitter someone asked, likely facetiously (I suppose that should be flippantly), whether functional MRI was next. FMRI of dolphins was the subject of an April Fool's Day joke a few years ago, and it does seem far-fetched at first blush. So, too, does studying trained dogs with fMRI, but Greg Berns' team is already doing that. Since thought experiments are cheap I figured I'd write a blog post to consider what might be feasible today if one were sufficiently motivated (read sufficiently well funded) to want to do fMRI of cetaceans. If nothing else we might learn something as we're forced to consider the manifold factors.


What's been done before?

Like any good experiment, thought or otherwise, we should begin with a review of the literature. (A month in the lab will frequently save an hour in the library, as the old aphorism has it.) And it turns out that some clever folks at UCSD have done both PET and SPECT scanning of a dolphin in the last few years. It's not fMRI but it is functional brain imaging and they used structural MRI (at 0.5 T) to locate the activity. Here's a dolphin getting a structural MRI scan in this figure from their 2006 paper:

(From Ridgway et al., 2006. DOI: 10.1242/​jeb.02348)

Structural MRI has a more extensive literature than the functional stuff. If you search PubMed for "dolphin brain MRI" you get 19 publications, most of which concern post mortem specimens. Some brains were fresh and in situ, others were fixed. But the papers from UCSD that reported PET and SPECT results seem to be the only published academic studies using MRI of a live dolphin, too. There are reports of dolphins receiving MRI for diagnostic medical purposes, such as this one, but they show CT scanners in their illustrations.


Sizing the hardware and getting a dolphin-friendly "patient bed"

I shall assume we want to be able to scan an adult bottlenose dolphin. The main constraints are the mass of the animal and the size and location of the dorsal fin and pectoral flippers. Open style (stand-up) magnets accommodate the dorsal fin easily but the magnet strength is low. Likewise, open MRIs of the type used by UCSD (shown above) accommodate the pectoral fins but again the magnetic field is low. The best I've found is the new 1.2 T Oasis from Hitachi. An impressive field strength for an open MRI, but let's aim for at least 3 T assuming we want lots of fMRI contrast and high resolution anatomical scans.

A wide bore 3 T MRI might work for us but before we commit to what's on the market we need to consider the dimensions and mass of an adult dolphin in more detail. The following dimensions were estimated from a paper on dolphin growth rate together with a basic view of dolphin anatomy courtesy of the Texas Marine Mammal Stranding Network. Our model adult bottlenose dolphin is 2.7 m long. The distance from the center of its brain to the front of the dorsal fin is about a meter. This distance is compatible with Verio & Skyra magnets from Siemens. The Skyra has 70 cm usable bore diameter with 173 cm total length, meaning that the center of the magnet is 86.5 cm from the front flange of the magnet. The leading edge of the dorsal fin would end up just outside the magnet.

But what about that 70 cm bore? For a length of 2.7 m the girth anterior to the dorsal fin is about 1.9 m. This gives a diameter at the widest point of 61 cm assuming our dolphin is (a) slim n' trim and (b) well approximated by a cylinder. We probably shouldn't assume either of these to be true. Furthermore, the entire bore isn't usable because we need some sort of table or other support to move the animal in and out of the magnet, and make sure it's not uncomfortable during the scan. While a flat table as for humans might work given a sufficiently large magnet bore, for a dolphin it's not the ideal shape. If our dolphin is 61 cm across at the widest point and the magnet bore is 70 cm diameter then we only have 9 cm to play with. An adult dolphin is considerably heavier than your average human, too, at 200-250 kg. We're going to have to replace the patient table. Would 9 cm be enough space to accommodate some sort of curved bed/sled that could move the animal in and out? Perhaps, with suitably strong materials or with the magnet bore lowered relative to the surrounding floor, as happens for some very large, high field magnets.

Cross section through a modern superconducting MRI magnet, showing all the principal components except the plastic covers that provide the external facade as well as some acoustic noise attenuation. From: Lvovsky, Stautner & Zhang, Supercond. Sci. Technol.26 093001 (2013). doi:10.1088/0953-2048/26/9/093001

Beyond supporting the mass of the animal there is an additional qualifier to the curved "dolphin bed." As I'll discuss more below, dolphins are highly sensitive to vibration and hear the majority of sounds through bone conduction via their lower jaw. Thus, serious consideration would be needed for anti-vibration measures when selecting materials for our bed/sled. It's difficult to fully decouple a bed from the magnet, even with cantilevered designs, so some sort of anti-vibration liner is probably implied.

What about gradients, RF transmission and reception? Here we can extrapolate reasonably well from what we use today for humans. As shown in the above illustration, a magnet-sized (so-called body) RF transmit coil and gradients are already accounted for in a commercial 70 cm clear bore system such as a Skyra. If it transpires that a larger clear bore is needed, e.g. to accommodate the custom bed/sled, then one obviously needs larger, custom gradient and body RF transmit coils proportional to the larger magnet bore. These wouldn't be cheap but the engineering is entirely tractable.

Signal reception could be achieved by something as simple as a blanket coil. The dolphin's brain is located just behind its blowhole, as shown below. We obviously don't want the animal to suffocate in the MRI so we would need to modify the blanket coil for breathing. Alternatively, one might easily pursue a custom coil design that resembles a large bowl and incorporate sufficient space for the blowhole. There are many, many options when it comes to RF signal reception, depending on whether you want to do accelerated image acquisition, how the coil is to be held in place, etc.

Basic anatomy of the Atlantic Bottlenose Dolphin. From the Texas Marine Mammal Stranding Network website.

Health and welfare issues

In addition to the vibration, what about sound pressure levels through the air? These are a big deal for humans, and dolphins are known to hear frequencies almost an order of magnitude higher than humans. Hearing is arguably a dolphin's most sensitive sense. If one were to design and build a custom 3 T magnet with a larger bore - say a usable bore of 1 meter diameter - then there would be sufficient space to install aggressive acoustic and vibration damping. In the absence of a more refined solution, one simple option would be to construct a fiberglass bore liner onto which several different densities of acoustic foam can be attached. (We did this for an old 4 T magnet years ago and the reduction in sound pressure level was better than 20 dB in the bore. Adding extra foam to the backs of the magnet covers - the facade you actually see from the outside - also reduces echoes and reduces acoustic noise both in the bore and outside.) Anyway, the bottom line on acoustic noise and vibration is that much designing and testing would be needed to ensure that the final sound/vibration level was safe.

At the same time it is important to ensure the shape and surface of the "dolphin bed" is comfortable and doesn't place undue pressure on any part of the animal's body. With humans we've used memory foam on the patient bed of my Siemens 3 T since we realized that people fidget less when they're comfortable. Memory foams allow better distribution of weight to minimize pressure points. For an animal used to having its weight supported by water this issue becomes even more critical, especially if one is expecting to conduct a scan over twenty minutes or so.

Keeping MRI subjects cool is always important, both for safety and comfort, but it is paramount for an animal that experiences external water cooling naturally. Blowing air down the bore is okay for (some) humans in an MRI but it will tend to dry out the dolphin's skin. The bore could be lined instead with a water misting system to keep the animal moist as well as cool. The magnet would need to be drained appropriately, and there would need to be suitable precautions to ensure water doesn't get where it's not wanted, such as into connectors for an RF coil. Would the mist or water pooling on the bed/sled cause issues for the MR images themselves, e.g. due to signal aliasing? Probably not if the right RF coil is used for signal reception. A full bore "body" transmission coil is already heavily (electrically) loaded with a 200+ kg dolphin inside it; a few liters of water isn't going to matter much. On reception, if we're using a curved surface coil like a bowl or even a blanket coil then we really only need to worry about water pooling on the coil's upper surface, a situation that could be ameliorated with drainage if the shape doesn't fix the problem for us. And of course we can always turn the misters off when we're actually scanning.




Physiology and BOLD

The goal here is functional MRI. Do we expect dolphin brains to exhibit BOLD contrast as we are used to seeing it in humans? On the one hand there is an abundance of evidence that mammalian brains exhibit rather consistent BOLD responses to stimuli. Studies of monkeys, dogs, cats and rodents have all shown robust BOLD contrast, whether awake or anesthetized. (Dead fish? Not so much.) We also have the UCSD studies using PET and SPECT as evidence that cerebral blood flow (CBF) might be estimated using the same methods in dolphins as for humans. So the chances of getting measurable BOLD and/or CBF changes in response to tasks would seem to be pretty good.

Cetaceans have evolved to dive and may be able to endure hypoxia for long periods compared to land-based mammals like us. Might this ability affect the BOLD contrast in interesting ways? Bottlenose dolphins generally have higher hematocrit and red blood cell levels than us land-lubber humans, unless we are high-altitude endurance athletes. And, just like us, the animal's fitness is a prime determinant of its blood constituents so we would want to record as many physiologic parameters as we can in order to distinguish neurovascular changes from concomitant effects. Variable hematocrit has already been implicated as an important experimental factor when studying exercising marine mammals. We're not going to have a dolphin diving in our MRI of course, but the dolphin's exercise history and general fitness may become crucial if we expect to compare fMRI results across multiple sessions. Definitely lots of factors to bear in mind when designing the fMRI experiment.

As for effects of the magnetic field, there is recent evidence that bottlenose dolphins can detect magnetism in otherwise identical lumps of metal concealed in opaque plastic barrels suspended in the water. But other than approaching barrels containing magnetic material more quickly, there was no difference in subsequent behavior: "The fact that all other behaviours did not differ between magnetized and control stimulus may reflect that magnetic fields are neither particularly attractive nor repulsive to dolphins."That's good. What we don't know, of course, is whether the strong magnetic field of an MRI would induce the dolphin equivalent of vertigo or other sensations that humans experience routinely.


Training for and conducting an fMRI experiment

I'm sure that most of you will have considered the ethics of attempting fMRI of a dolphin well before this point. I have neglected it until now simply because I wanted to consider the practical hurdles first. If we had run into an insurmountable problem then the ethics, while interesting, become moot. At the point of wanting to conduct an actual experiment, however, the ethics of the enterprise loom large and unavoidable.

Let's set a few limits. I'll assume that we want to scan animals that are available only because of extenuating circumstances. We are not going to attempt to scan wild dolphins unless by some miracle a wild dolphin can be coaxed into a scanner and lay still for the duration without significant assistance. Seems unlikely, even if an MRI scanner were partly submersed off a dock in Santa Cruz and a dolphin could swim right up into it. So we are talking about attempting to use dolphins in captivity only. I don't particularly relish the thought of dolphins in captivity for any reason, but I do understand that there are times when dolphins end up there and it may be inappropriate to release them. For example, some animals may have been rescued as orphaned youngsters or born into captivity and so lack the survival skills of an adult wild dolphin. Others may have been injured to the point where survival in the wild is unlikely. No doubt this would be the most contentious aspect of doing fMRI of a dolphin so I'll leave the ethics for you to debate in the comments below.

For the purposes of this thought experiment I shall assume that we have an adult dolphin living in captivity through no fault of its own. Furthermore, I shall assume that a committee of suitably qualified folks has determined that the perceived benefits of scanning the animal are sufficient to overcome objections and risks, just as we decide when studying any vulnerable population. As we move on to consider the training required for a dolphin fMRI experiment we will need to think carefully about the types of stimuli and responses we might want to use and determine the limitations imposed.

Perhaps the simplest fMRI experiment would be the "resting state." Here we might conceive of training a dolphin to lie still in a tube for ten minutes in exchange for a reward, such as Craig Bennett's dead salmon. This might permit discovery of the intrinsic connectivity in the dolphin's brain. Task-based fMRI would involve a far higher degree of training, but with a suitable mock scanner and enough fish rewards there would be a good chance of success if what's already been achieved with dogs is any guide.

In most human fMRI studies, and in the earliest dog fMRI studies, the stimuli are visual. Is that a good choice for a dolphin? Since I'm not a marine mammal expert I have no idea. What I do know, however, is that a dolphin's eyes are more on the sides of its head than forward-looking, as for humans and dogs. If we are going to use visual stimuli then we need to be thinking about the binocular visual abilities of the dolphin, perhaps resorting to twin goggles for monocular or binocular display, or a single target mounted sufficiently far from the dolphin's nose for it to see with both eyes.

Other stimuli might be possible. Those clever people at UCSD came up with a device that permits a dolphin to echolocate while out of water. The echolocation clicks "produced by the dolphin are detected with a hydrophone embedded in a suction cup on the melon." For stimuli: "The experimental approach relies on the generation of electronic, or “phantom” echoes, rather than the use of a physical target directly ensonified by the animal." How does the dolphin sense the echoes? "The echo signals are then projected to the dolphin via a transducer placed on the lower jaw." Here's a schematic of the echolocation pathway for a dolphin, taken from the Wikipedia page on animal echolocation:

Sound generation, propagation and reception in a toothed whale. From the Wikipedia page on animal echolocation.

That would surely be the hard part done, especially as the melon is forward of the brain and blowhole. When mounted the device ought not be in conflict with a receive RF coil. Making the device MR compatible ought to be straightforward by comparison to inventing it in the first place! Would it be possible to use this device during scanning while ensuring good attenuation of the scanner's vibrations and noises? The scanner peak acoustic frequencies are typically less than 3 kHz whereas dolphins echolocate from 40-150 kHz. Even allowing for a few harmonics there should be minimal energy produced by the scanner in that range.

The whistles and clicks that humans can hear are obviously at much lower frequency than those used in echolocation. Could we use the audible vocalizations as our dolphin version of a button response box? Perhaps. The interference of scanner sounds is an issue but again, with good acoustic engineering and careful tuning of the scanner peak frequencies I don't see why this couldn't be achieved, provided the vocalizations don't require movement of the head or body. These are issues that those doing human fMRI of audition and speech already deal with.



Functional MRI of dolphins would seem to be a reasonable experiment to pursue if one has sufficient resources and the freedom to perform extensive customization of a scanner. I don't foresee any major technical impediments even if the logistics would be a tad involved. And, as they are the second most intelligent species on this planet, after the mice, I think it would be fascinating to see a dolphin's brain in action. Who knows, perhaps with fMRI we will be able to decode the dolphins' warnings of our imminent destruction by the Vogons. Or we could try to find a Babel fish to shove in an ear. Choices, choices.

This is going to be among the more unusual blog posts I've written. Here goes nothing.

The back story. A few years ago I decided to take a stab at writing fiction. The result was Bubble Chamber, an academic satire with a scientific twist. My intention? To see if I could write a novel, to actually finish it and get it into the public domain as an end point. Check! The process wasn't even sufficiently painful to dissuade me from future forays into novel writing. Just not yet, I have too many other irons in the old fire.

Towards the bottom of the post you can read a synopsis of Bubble Chamber along with a little contrived biography of yours truly. I am also tempted to include here a link to Amazon so that you might procure a copy of my literary scribbling for yourself. But I won't do that just yet. We need to make a quick detour.

Dementia Adventure

It transpires that a young Englishman by the name of Chris Graham has been steadily riding his bicycle, solo and unsupported, around the circumference of Canada and the USA since the beginning of May. He is now in Maine, approaching Canada for the second time and about to commence the last leg of his marathon journey. His current location, as well as the route already completed, can be seen in real time courtesy of his YBTracking page. A flight back home to the UK on Christmas Eve beckons as the final reward for all of Chris's hard work. (Motto: If in doubt, pedal!)

Chris is doing his ride for charity. Specifically, Chris is raising money for Alzheimer's Research UK. You see, Chris is likely to develop dementia in a not too distant future. He carries the gene for early onset Alzheimer's, a disease that has claimed the lives of many of his close relatives, including is father and grandfather, at an age that Chris himself is now rapidly approaching. You can learn more about Chris and his family history on his web page.

This is no sob story, however. Far from it. In fact, following Chris on his Facebook page has brought smiles and laughs to many thousands of people this year. Watch one of his video updates then tell me you won't be back to see what happens next! Better yet, quit reading this blog and pop on over to Chris's Just Giving web page to send him some money right now, then go peruse the photos and videos on his Facebook page. Thanks, and goodbye!



If I didn't just lose you to an immediate donation then I have a proposition for you. Chris is a mere three lousy UK grand from hitting a £40K target and I want to do a little something to help him get there. All proceeds arising out of sales of my aforementioned novel were committed 100% to charity from the moment I conceived of writing it. That hasn't and won't change. An assisted living charity in the eastern US has received very occasional, very small checks over the years. But for the rest of this year at least, all proceeds from sales of Bubble Chamber - available via this Amazon page - will go to Dementia Adventure. I am also going to send along an equivalent amount of my own money for each copy sold so that you have extra inducement to go buy the book.

If you prefer to ignore my book entirely and simply donate to Chris that is just peachy. Or donate to Chris and then go buy the book, I won't stop you. Either way, join Chris virtually as he goes into the last three weeks of a truly epic bike ride. Along the way, take the time to think about just what it is you want to achieve with what remains of your life. We could probably all use a little bit more of Chris Graham's irrepressible spirit.

 

Bubble Chamber - a novel

Simon Mills had never actually planned to become a scientist. But he hadn't managed to avoid the eventuality either. As with most things in his life, Simon's drift into academia just sort of happened, the result of a series of coincidences and twists of fate over which he seemed to exert minimal control, regardless of whatever alternative path he might have preferred. 

So, instead of spending his early twenties enjoying a slow, predictable existence in the English countryside of his youth, Simon finds himself first as a reluctant post-graduate student in London, working for a quintessentially eccentric professor, and then - and much to his chagrin - propelled across the Atlantic to the USA, to pursue some harebrained experiment or other. All in the name of science. 

Simon lurches from one unsolicited career opportunity to another, finally getting himself caught up in an academic political battle going on half a world, and a whole scientific dimension, away from his own. Not that distance or relevant experience appear to be factors when it comes to Simon Mills' vocation. It seems fortune just won't leave him alone.

From the Publisher

About the Author

SCH Thurston is a real scientist at a real university somewhere in America. When he isn't fighting academic political battles or trying to balance declining budgets, he can be found dabbling in a bit of actual research.

Thurston is in the process of writing a second novel. Another academic satire, "Lem" is set in a regional university in small-town Australia. Without wanting to give too much away, the story blends space exploration with ecology and the 'green' zeitgeist, throws in a critique of public relations and media manipulation for good measure, and ends up with an utterly ridiculous take on how science is portrayed to the public at large. It is a satirical look at the interplay between journalists in the mainstream media and scientists in academic research. In Thurston's ill-considered opinion, both sides could probably do better.

(Note: "Lem" is on indefinite hold after 50,000 words. Other projects and international events conspired to render a major plot rethink essential. One day.)

(A free copy of Bubble Chamber to anyone who can guess how my nom de plume came about.)

Most labs have plastic goggles for correcting a subject's vision in the MRI. Here's our box of tricks:



These goggles work pretty well for a standard sized head coil, such as the 12-channel TIM coil on my Siemens Trio. But for a tighter fitting coil, such as the 32-channel head coil, there is simply no way to cram a subject wearing goggles into the space available. For a start the goggles' frame prevents the subject's nose from penetrating the appropriate gap in the front of the coil.

A simple solution is to relocate the corrective lenses on the outer surface of the head coil. All that's required is a different way to hold the lenses in place. My ace engineer, Rick, made this pair of holders for our 32-channel coil:



The plastic strip with the label on it is actually optional, it's a way to keep the otherwise independent lens holders together in the store cupboard. Here's what one of the holders looks like with lens inserted:



The plastic mounts into the gap in the RF coil with moderate friction to hold it in place. In the above photo you can also see the black velcro strips that can be used to hold the pair of holders together, but as mentioned above the holders actually mount independently.

Below are a few more quick pix of a holder showing its basic construction: a block cut to the right size to fit snugly into the socket on the coil, and a smooth groove to hold the lens in place. I have no idea if there is a company making these things but they are basic enough that you could fashion some yourself in an afternoon.

Several people have approached me for advice on using simultaneous multi-slice (SMS) EPI for fMRI experiments. This is the sequence also known as multiband (MB) EPI. I'll come back to nomenclature in a moment. First, though, a brief introduction to what may become a lengthy series of posts. I'm going to focus on BOLD-based fMRI exclusively for the time being - sorry diffusion and ASL folks - and because I presently only have a Siemens Trio at my disposal, everything I write will have strong bias in that direction. That said, I do anticipate writing later posts dealing with SMS-EPI (for fMRI) on a Siemens Prisma at least, and I can already envisage a need for posts dealing with receive field normalization, in-plane parallel imaging, distortion correction options, reconstruction options and multi-echo SMS, to name just a few advanced topics. But first things first - to get going!


Options for SMS-EPI on a Siemens 3 T scanner

I am aware of three SMS-EPI pulse sequences for a Siemens Trio. One comes from the University of Minnesota's Center for Magnetic Resonance Research (hereafter CMRR), one comes from the Martinos Center for Biomedical Imaging at Massachusetts General Hospital (hereafter MGH), and one comes from Siemens as a work-in-progress (WIP) aftermarket sequence. For this post I'm going to be using the sequence provided by CMRR. Since CMRR refer to their sequence as multiband (MB) EPI I shall stick to this nomenclature here, and reserve the term SMS-EPI to apply to the broader family of pulse sequences. I may do posts on the MGH and WIP sequences in the future, but the CMRR sequence has been used the most broadly to date (e.g. the Human Connectome Project, which I'll discuss at length below) and so it offers the most immediate, road-tested place to start.

Other pertinent background. In addition to being on a Siemens Trio, I am going to assume you have a 32-channel receive-only head coil. The MB-EPI and other SMS-EPI sequences can be made to work with a 12-channel coil but only in a much reduced fashion because the 12-channel coil is a Total Imaging Matrix (TIM) coil that generates a lower number of effective channels than the headline number suggests. Specifically, the 12-channel TIM head coil acts like a 4-channel coil (in Triple mode) and that means it has limited RF heterogeneity with which to encode simultaneous slice information. Furthermore, the geometry of the 12-channel coil isn't ideal because it comprises nearly parallel, only slightly curved struts that have very little spatial variation along the coil axis, which happens to be the axis most people want to slice along (for axial images). There's a comparison of my 12-channel to 32-channel coils in an earlier post on receive field heterogeneity as an artifact rather than a feature.

Your first task is to obtain the CMRR pulse sequence, for which you'll need a master research agreement (MRA) with Siemens and then a C2P agreement with CMRR. Not 100% sure what C2P stands for but I believe it's Core Competence Partnership. In any event, once you have the sequence in-hand you should of course read the release notes and review their gallery of images before you do anything more. I would also suggest that you or your facility physicist spend a couple of hours perusing the history of the current and closed issues in their GitHub. You may well find some of your technical questions already answered. And given that this is a relatively high risk experiment compared to single shot EPI, you or someone around you is going to want to have a decent amount of expertise testing on phantoms and volunteers before you begin using SMS-EPI for an entire study.


Why SMS-EPI at all?

For most fMRI researchers, SMS-EPI is likely to be of interest when you want voxels with dimensions smaller than about 2.5 mm on a side, and/or you want a repetition time (TR) substantially below 2 seconds while maintaining at or near whole brain coverage. There are several published studies showing advantages of SMS-EPI for fMRI. A PubMed search for "multiband EPI" and then "SMS EPI" will locate them. Add "BOLD" as an additional search term if you want. Advantages of SMS-EPI versus regular multislice EPI may include better characterization of physiologic noise, leading to better discrimination of resting state networks, higher spatial resolution without losing whole brain coverage, and possibly higher sensitivity to BOLD changes. I'll leave the motivation to you, but you will definitely want to review all the published studies relevant to your neuroscience. If in doubt, keep it simple!


What the Human Connectome Project recommends

The HCP team wrote two papers containing essential information on their fMRI protocol. I'm going to review the salient parts of these papers before I get into their specific recommendations because what they did for HCP is actually different than what they recommend. Why? Because they had a customized 3 T scanner - the "Connectome Skyra," with more capable gradients than any stock 3 T.

In the paper describing the resting-state fMRI portion of HCP (doi:10.1016/j.neuroimage.2013.05.039) they explain how they settled upon 2 mm isotropic voxels, noting:

Pilot studies using a range of resolutions and EPI accelerations indicated that reducing voxel size to less than 2 mm was not beneficial at 3 T, in terms of the spatial detail discernible in the rfMRI resting-state correlation structure. This was not only due to the reduction in SNR with increased spatial resolution, but also the relatively large point spread function of the BOLD effect at 3 T (Parkes et al., 2005), which was measured to be ~ 3.5 mm at full-width-at-half-maximum (FWHM) caused by the dominant draining vein contribution at this field strength (Uludağ et al., 2009).

In the finalised protocol this acceleration is used to acquire 2 × 2 × 2 mm data with a temporal resolution of 0.72 s. While a larger voxel size would result in even faster imaging and better SNR, this choice provides a good overall balance in which both spatial and temporal resolution are a significant improvement compared with conventional rfMRI datasets.

Other key parameters followed from the voxel size and consideration of signal losses relative to BOLD sensitivity:

Echo-time (TE) was, after much discussion and evaluation, set to 33 ms. Again, this choice is a trade-off; long TE increases BOLD contrast, but decreases overall signal level and increases signal dropout in areas of B₀ inhomogeneity. The TE for optimal functional CNR is equal to T₂* when thermal noise dominates; however, T₂* varies spatially, meaning that no single TE can be optimal throughout the entire brain. Acquiring multiple echoes in a single EPI readout train, or in separate acquisitions with multiple TEs, was not acceptable due to significantly prolonged readout duration and/or TR. Thus, the shortest TE that could be achieved without the use of partial Fourier or in-plane accelerations was selected, to minimise signal dropout. At 2 mm resolution, with the Connectome scanner gradients, this TE was 33 ms, given the excitation pulse width (~ 7 ms) required to achieve the (Ernst) flip angle (52°) for multiband x8. The use of partial Fourier to reduce TE resulted in larger signal dropouts than acquiring a full Fourier coverage of k-space with longer TEs (likely caused by local phase ramps in regions of B₀ inhomogeneity shifting signal outside the acquired k-space region); hence partial Fourier was not utilised. The EPI echo train length is 52.2 ms for the final HCP fMRI protocol.

These are useful points to keep in mind as you proceed. If you want to review all the MB-EPI parameters as used in the HCP, check out pages 41-3 of the PDF accompanying their release of data from 900 subjects. The HCP team also wrote a paper on their minimal processing pipeline (doi:10.1016/j.neuroimage.2013.04.127) which you will want to read before setting up your own experiment, but we don't need to dig into that just yet because, as I mentioned above, what the HCP recommends we do is different to their setup.

So, just what do the HCP team suggest we do on a standard 3 T scanner such as my Trio? As noted above, we don't have a high-performance custom gradient set as on the "Connectome Skyra" so we have a few more limitations to deal with. (Perhaps those of you with a Prisma will be able to push things harder. I'll deal with Prisma-specific protocols in a later post.)

For the HCP recommendations we need to review pages 3-7 of that PDF accompanying the 900 subjects data release:

For functional imaging, key choice points relative to the HCP fMRI acquisitions involve the multiband (MB) factor, spatial resolution, TE, and phase encoding direction (the latter three of which all interact). While gradient strength is not as critical for fMRI (relative to dMRI), the Connetome Skyra gradients do allow it to operate at a lower echo spacing than a conventional 3T scanner (e.g., 0.58 ms vs. 0.69 ms at 2 mm, all other things being approximately equal). The limitations of maximal readout gradient [Siemens Trio (TQ) ~ 28mT/m, Verio (VQ) and Skyra (XQ) ~ 24 mT/m] and forbidden echo spacing (due to acoustic resonances) make 2 mm more of a “stretch” resolution on these 3T magnets.

More of a stretch? Yes, sadly that's not just a figure of speech. The echo train length will necessarily increase on a scanner with weaker gradients. We'll deal with this issue for the in-plane optimization, below. Until then, here's what the HCP summary recommendations look like:

Note that considerable benefit as regards the accuracy of the mapping of activation to the cortical surface is already achieved by going to a 2.5 mm isotropic resolution, albeit with further incremental gains in accuracy in going down to 2.0 mm (Glasser et al. 2013). Overall then, we recommend that users of Trio, Verio, and Skyra systems test resolutions of 2.0 to 2.5 mm for fMRI and make a selection based on their requirements for temporal SNR, statistical power, and acceptable degree of susceptibility distortion, and signal dropout.

The good temporal stability of the Connectome Skyra and the low electronic noise of the Siemens Tim 4G© platform allow the HCP to robustly generate good quality BOLD data at an MB factor of 8 without in-plane acceleration. Users of other systems will want to look carefully at whether they are happy with the levels of residual aliasing and temporal SNR at MB=8. In general, we recommend a MB factor of MB=6 for robust image quality while retaining high temporal resolution for these systems.

There's one final piece of advice concerning the phase encode direction. HCP opted to use a rectangular field-of-view, with phase encoding alternating L-R and R-L. But they don't suggest we do this:

....we recommend using either anterior-to-posterior (AP) or posterior-to-anterior (PA) phase encoding (rather than the RL and LR phase-encoded pairs used in the HCP acquisitions), so that there is not a right/left susceptibility asymmetry (bias) in the aggregate data. In pilot testing, we could not discern an overall preference for either AP or PA phase encoding, since each resulted in a different amount of signal dropout and local distortions in different brain areas with susceptibility artifact, and this dropout differs greatly depending on slice orientation (e.g., T>C vs C>T). Thus, we recommend that users make the choice between AP and PA phase encoding based on their own particular research aims and goals. Note that AP or PA phase encoding will require use of a full FOV in the phase direction (“FOV phase = 100%”), which will lengthen the total echo train, leading to some increase in T2* blurring, susceptibility distortion, and signal dropout (via increased minimum TE) compared to the HCP acquisitions. In practice, this effect will be at least partially mitigated given the shorter minimum echo spacing achievable in the AP/PA phase encoding direction (due to lower peripheral nerve stimulation limitations with AP/PA than RL/LR). “Compensating” for these effects via use of partial Fourier and/or in-plane GRAPPA involve their own tradeoffs (e.g., for in-plane GRAPPA, reduced image SNR and a lower acceptable maximum multiband acceleration factor and thus longer minimum TR). The HCP investigated these tradeoffs to some degree during pilot testing, and ultimately settled on RL/LR phase encoding with no partial Fourier or in-plane GRAPPA as yielding the best overall quality on the Connectome Skyra. For users of other Siemen’s 3T systems desiring 2.3 – 2.5 mm isotropic spatial resolution with only a single phase encoding direction, we suggest trying AP or PA phase encoding without in-plane GRAPPA or partial Fourier (allows a minimum TE of ~ 33 ms), and a multiband factor of 6. For users desiring 2.0 mm resolution, 7/8 partial Fourier may be desirable (allows a minimum TE of ~ 36 ms).

These suggestions give us sufficient information to proceed. To make this a tractable project I shall first describe in general terms the main factors to consider, then put all the decisions together into three explicit starting point protocols.


Slice dimension optimization

To a first approximation, the amount of slice acceleration - what is generally called the MB or SMS factor - depends upon the RF coil selection. As noted, we are assuming a 32-channel head coil. Not only does this coil have more independent channels than the standard 12-channel coil, but the elements also have very different geometry. In the 12-channel coil the elements are arranged as near parallel struts; there is little receive field variation along the coil's length. The 32-channel coil's elements are arranged more like the hexagonal patches on a soccer ball. There is a high degree of receive field anisotropy parallel as well as perpendicular to the coil axis. Thus, in principle we might choose to slice along any axis when using the 32-channel coil.

So, what direction should we place our slices? Efficiency dictates that we cover the shortest brain dimension with the slowest sampling. Since slices acquire more slowly than either of the in-plane dimensions, even with SMS enabled, it means that we should expect to use axial or axial-oblique slices. The 32-channel coil has plenty of receive heterogeneity along the axial direction (magnet Z axis) to perform good MB-EPI. You could use AC-PC or some other oblique axis equally well.

How fast should we go? The HCP suggests that we set MB as high as 6. This results in good images on my Trio and so this has become my practical limit, too. Would I consider going to, say, 8? I might, but it's worth bearing in mind that an MB of 6 is already six times faster than regular multislice EPI. That's a phenomenal gain! An MB of 8 is only 33% better than an MB of 6. And it's higher risk. Conversely, if an MB of 6 nets all the brain coverage you want with a minimum TR that is much shorter than you require, e.g. to use a particular task script that triggers off the TTL pulses happening once/TR, then you should consider dropping the MB to 4 or even 3 to reduce the possibility of reconstruction artifacts, and to decrease the image reconstruction time. (Vide infra.) In sum, then, accelerate as much as you need to, but not more!


In-plane optimization

You might think that in-plane optimization should be trivial because it's the slice dimension being accelerated in SMS-EPI. In practice, however, the separation of the simultaneously acquired slices needs some help from a slight modification to the in-plane gradients, via a method called blipped CAIPI. Reconstruction artifacts - say we try to push the acceleration factor higher than can be supported by the RF coil - may appear in the image plane, not just as modulations only visible in the slice dimension. You may see these artifacts described as "residual aliasing" or, more commonly for the SMS literature, as "leakage artifacts." I'll look at these in detail in a subsequent post.

There is also a prosaic reason why we shall want to change the parameters in-plane. When the slice thickness goes down there is a tendency to want to improve the in-plane resolution simultaneously. Many (most?) people seek voxels having nominally isotropic dimensions - cubes - or very close to it. What this means is that because we have enabled higher resolution in the slice dimension we're now motivated to drive the resolution in-plane. There's no obligation to do so, it is a choice but it's one that many of you will succumb to.

Following the lead of the HCP recommendations, there are a few guiding principles that we shall use for our starting point protocols. Unfortunately, high resolution in-plane takes gradient time and so the minimum TE will end up being longer for high in-plane resolution than for the 3 to 3.5 mm voxels you're used to. Our guide will be to try to set TE as close to 30 ms as possible, but not longer than 40 ms, in order to match the typical T2* of most of the brain. We should expect regions prone to signal dropout - the frontal and temporal lobes - to suffer more the longer TE gets.

We generally want to drive the echo spacing as short as it will go, while noting that some scanners have mechanical resonances that can make ghosting bad. Newer Siemens operating software - certainly the SyngoMR VD line as used on Skyra, for example - generally prohibits the use of an echo spacing known to cause mechanical resonances. But on my Trio I can set the gradients to any physically plausible value and thus I need to be aware of the consequences.

Finally, use of GRAPPA in-plane is not yet encouraged by me (or HCP) unless you take special measures to ensure that head/body motion is considerably lower than used in most routine fMRI studies. All the subject has to do is swallow during the autocalibration scans (ACS) and a whole time series can be wrecked. Also, in agreement with the conclusions of the HCP team, I've noticed that the raw image SNR is far lower when GRAPPA is enabled. So no GRAPPA for now. There are a few recent options for ACS that may make GRAPPA sufficiently robust we might want to consider it in future, so I'll do a dedicated post on using GRAPPA with SMS-EPI soon. Until then, if you insist on enabling GRAPPA with your SMS-EPI then you should probably take your cue from this paper. (Note, however, that GRAPPA was always enabled in their comparison, no non-GRAPPA option to compare against.)


Dealing with dropout

For voxels smaller than 2 mm on a side we shall necessarily have a TE in excess of 30 ms, causing more dropout in frontal and temporal regions than you would like. While the HCP was able to avoid using partial Fourier courtesy of their zippy custom gradient set, for us it is going to be a useful trick. It's not without consequences, but we have a degree of control over their severity by setting our partial Fourier fraction in conjunction with the direction of the phase encoding; either A-P or P-A. If you aren't familiar with partial Fourier, especially the interaction of partial Fourier with the phase encoding direction, then I strongly suggest you read the three posts linked in this paragraph. For MB-EPI we shall be using the Siemens default "early" echo omission, in order to reduce the minimum TE. (There is a "late" option but it's not useful for us when we want to shorten TE.)


Dealing with distortion

There are generally two reasons to want to use in-plane acceleration such as GRAPPA. One reason is to try to shorten the TE, the other reason is to reduce the amount of distortion in the phase encoding direction. We are already expecting to employ partial Fourier to reduce TE below 40 ms for MB-EPI, so that leaves the latter reason: distortion reduction. Now, we are going to have severely distorted MB-EPI images because we are expecting to employ high resolution, and high resolution in EPI means long echo trains. Long echo trains produce severely distorted EPIs. Sorry. This is one of those situations where we have to choose between two unpalatable options: motion sensitivity or distortion. For now I choose distortion because there is some hope to mitigate the worst effects with a post-processing step employing a field map, or functional equivalent. Motion, on the other hand, tends to trash a time series irrevocably.

So let's take a quick look at what the HCP did for distortion in their protocol:

To more rapidly measure the _B0 field for correction EPI distortions, we acquire two spin echo EPI images with reversed phase encoding directions (60 s total for 3 pairs of images). (Note that we refer to these images as spin echo field maps, though they measure the field by reversing the phase encoding direction, which is a very different mechanism from standard field maps that use a phase difference calculated from two different TEs.) These spin echo EPI images have the same geometrical, echo spacing (0.58 ms in our scans), and phase encoding direction parameters as the gradient echo fMRI scans. These images enable accurate correction for spatial distortions in the fMRI images so that they can be precisely aligned with the structural images. Two of these spin echo EPI field mapping pairs are acquired in each functional session, for added robustness with respect to acquisition errors and subject movement, along with one set of B₁− receive field-mapping images (with identical parameters to those described in the structural session).

For the functional pipelines, a field map is required, because any neuroimaging analysis that aims for precise cross-modal registration between functional and structural (or other data modalities) will require EPI distortion correction. In general, either standard gradient echo field maps or spin echo EPI field maps can be acquired, though spin echo EPI field maps can be acquired more quickly with less chance of motion corruption. The geometrical parameters (FOV, matrix, phase encoding direction, resolution, number of slices) and echo spacing must be matched between the gradient echo EPI fMRI timeseries and the spin echo EPI field maps.

EPI fMRI images contain significant distortion in the phase encoding direction. This distortion can be corrected with either a regular (gradient-echo) field map or a pair of spin echo EPI scans with opposite phase encoding directions. Reversing the phase encoding direction reverses the direction of the B₀ field inhomogeneity effect on the images.

There are other options, such as the true (gradient-echo) field map mentioned by HCP. Siemens offers a sequence called gre_field_mapping and released as part of their BOLD software package. (Check the SIEMENS tree in the Exam Explorer.) But on my scanner at least, the minimum slice thickness is 2 mm and you may wish to match exactly the field map with the MB-EPI slices when the latter are thinner than 2 mm.

The issues of acquisition time, motion sensitivity, correction efficacy and so on make this a major enterprise and I don't have any quick suggestions for you, so I shall leave this extensive topic for a dedicated post once I have more details to offer. Until then I suggest you take a look at what Oregon's LCNI recommends, and perhaps consider using FMRIB's FUGUE software, in addition to considering HCP's suggestions.


Other issues

Shimming:  The CMRR-supplied protocols come with Advanced shim enabled and that should be considered the default. It takes about 90 seconds but the benefits in frontal and temporal lobes should be worth the extra time. If you're looking to nick a few minutes in a lengthy scan session then you might consider starting with the Advanced shim for the first fMRI run, then use Standard shimming for all subsequent runs.

Flip angle:  The HCP used an approximate Ernst angle for their TR; 52 degrees for TR = 720 ms. I tend to use a lower FA than the Ernst angle for two reasons. First, as shown in this paper from Gonzalez-Castillo et al. (doi:10.1016/j.neuroimage.2012.10.076), reducing the FA can reduce physiologic noise without reducing BOLD sensitivity. (Years ago on a 4 T we always used a very low FA, such as 20 degrees for a TR of 1-2 seconds, in order to reduce the inflow effects. It worked very well.) The second reason might be less of an option. The amount of subject heating (assessed by the specific absorption rate, SAR) scales as the square of the FA. So even a small reduction in FA can reduce SAR considerably. If you find yourself running into the SAR limit then you can reduce the FA until you're good to go. The down side to a lower than Ernst angle FA? Image contrast may be affected slightly, but since we prefer temporal stability over anatomical content then altered contrast shouldn't be a huge burden. (For Gonzalez-Castillo et al. it was an advantage because they were using water-excite rather than fat presaturation on a GE scanner and so they had low anatomical contrast at the Ernst angle.)

RF pulse width:  The HCP used an RF pulse width of 7120 us whereas the default on my protocols is set to the maximum of 10500 us. I haven't tried altering the RF duration at all yet, and I will note that I can get down to 1.5 mm slices using the default which is why I've not played with it. The "Connectome Skyra," with its higher peak gradient capabilities, could probably get thin slices with the shorter duration RF. On our stock scanners we should expect to need longer pulses. So until someone shows me why 10500 us is a bad choice I plan on continuing with it.

Receive coil heterogeneity:  According to the PDF accompanying data release, the HCP acquired receive field "bias" maps but actually used a different approach:
Having looked at both the PDF and Glasser et al. it's still not entirely clear to me whether they were using the T1W and T2W scans or the two named BIAS scans. Perhaps someone can enlighten me. In any case, I advocate enabling the "Prescan Normalize" option on MB-EPI provided you also select the option to save both the normalized and raw time series (enable "Unfiltered images"). Reconstruction time is dominated by the SMS unaliasing algorithm, not the normalization, and so the only significant cost to having two data sets is twice the amount of data to store. More on bias field correction in a later post, after I've had a chance to assess the pros and cons of the Siemens routine versus acquiring one's own reference data set.

Coil combine mode:  I would only use the default Sum of Squares for simplicity, as the HCP did. I don't know enough about the alternative methods yet, but my intuition is that fancier coil combination methods may not play nicely with the SMS unaliasing routine. 

Single band images (SBref):  The HCP used the single band references as their template for realignment (a.k.a. motion correction) and to permit more accurate registration of EPI to T1W anatomical scan. I can't say whether the SBref makes a better or worse template for motion correction than the first MB-EPI, the middle MB-EPI, or some average. This is a post-processing question that we can ignore at acquisition time, provided we ensure that we have the SBref saved along with the MB-EPI time series.

Reconstruction speed:  So this will likely be your loudest complaint when first using SMS-EPI. I am reliably informed that the GPU-based reconstruction engines on the newer scanners can process SMS-EPI in real time. Not so the Step IV MRIR board on my Trio, and this is the fastest reconstruction board you can get for the Trio. That's why it's useful not to set the MB factor to some gratuitously high number for the sake of having more data. More may or may not be better, but it will most definitely be slower. With MB=6 you can expect the reconstruction to run longer than the acquisition by between 1.5 and 3 times. Plan your session accordingly. Also note that a maximum of twelve reconstructions can be in the Siemens queue at one time. I've never hit that limit, or even come close. What tends to be a bigger deal is the need to keep the next user from terminating a reconstruction at the end of the current session.

Slice order:   Simultaneously excited slices already have a sizable gap between them such that (a) we don't need to specify additional gap, and (b) we can reconsider the desire to use contiguous slices for reducing motion sensitivity, as is common for regular single shot EPI. For an MB of 6 there will be a gap of five slice thicknesses between the simultaneous slices. For ordering, there is a difference between specifying interleaved slice order versus contiguous (ascending/descending) order, but it isn't a stark difference as for regular EPI. So for convenience I have stuck with interleaved slices. I doubt there is much practical difference if one selects a contiguous order, but that would necessitate me checking for artifacts all over again. I don't expect there to be an appreciable difference in motion sensitivity, for example.

Motion sensitivity:  I can only offer an anecdotal report on the motion sensitivity of MB-EPI. In my experience, moderate subject movement (a swallow, say) during acquisition of the single band reference images has a much smaller effect on image artifact level than the same motion would during ACS for in-plane GRAPPA. This is good! The HCP did some piloting and assessed the ability of ICA to clean the time series, but I suggest that you treat MB-EPI (and all SMS-EPI sequences) as intrinsically more motion-sensitive than regular EPI, and adapt your subject preparations and head restraint systems accordingly.


Putting it all together

I've created three introductory protocols for MB-EPI based on three different voxel sizes: 1.5 mm, 2.0 mm and 2.5 mm dimensions, all isotropic:


The full parameters are available in a PDF, a printout of the full protocol from the Exam Explorer. The protocol was created using CMRR's MB-EPI sequence version R012c (R013 is out as of 14th December, 2015) on my Trio running SyngoMR VB17A. There are two sets of scan parameters for each voxel size in the protocol, one with phase encoding set A-P and the other set to P-A. Phase encoding reversal is achieved via the parameter "Invert RO/PE polarity" on the Sequence tab. Enabling this option is functionally equivalent to setting the "Phase enc. dir" dialog box to 180 (degrees), which is why you still see A-P listed under that parameter. All six scans have prescan normalization enabled, SBref saved and excite (RF) pulse duration set at 10500 us.

There are some caveats that require explanation. The 1.5 mm isotropic voxel protocol doesn't cover the whole brain in the TR of 1300 ms. It might just cover whole cortex with judicious slice tilting to snag temporal poles as well as the full frontal and parietal cortices. You'd need to extend the TR out to about 2000 ms to include cerebellum. You might consider setting MB up to 8 to improve the coverage, but I can't vouch for the leakage artifacts that might result.

For the 2.5 mm isotropic voxel protocol the whole brain is covered easily with 64 slices in a TR of 1200 ms. If you wanted an even shorter TR, down to 800 ms, you might consider setting the MB factor to 6 rather than 4.

The 2 mm isotropic voxel protocol uses an echo spacing of 0.69 ms, right in the middle of the range 0.6-0.8 ms where mechanical resonances for axial slices are expected to be worst on my Trio. Unfortunately, there is insufficient gradient strength to get below 0.6 ms, and setting the echo spacing at 0.8 ms would necessitate a TE above 40 ms, thereby requiring 6/8ths partial Fourier to get back below 40 ms. In my throwaway tests the ghosting from the echo spacing of 0.69 ms weren't terrible (see the example data below) whereas the dropout from using a longer TE and/or higher fraction for partial Fourier produced noticeable dropout. You would want to verify on your scanner that such an echo spacing doesn't produce show-stopping ghosts; test on a phantom. The mechanical resonances differ from installation to installation.


Example data

Below are 3D-MPR displays of a single volume of MB-EPI data acquired from the same brain using the three protocols just outlined. Here the "Prescan Normalize" filter has been applied. I also have 50-volume time series acquisitions with and without "Prescan Normalize" that you can download and tinker with yourself (295 MB zipped file). Included in the file are the "split" mosaic images I used to create the 3D-MPR views.

2.5 mm isotropic voxels (click to enlarge)

2 mm isotropic voxels (click to enlarge)

1.5 mm isotropic voxels (click to enlarge)

References and further information:

Human Connectome Project:
Appendix I – Protocol Guidance and HCP  Session Protocols
(900 Subjects Release, 8th Dec 2015)

The minimal preprocessing pipelines for the Human Connectome Project
MF Glasser et al., NeuroImage80, 105-24 (2013)
doi:10.1016/j.neuroimage.2013.04.127

Resting-state fMRI in the Human Connectome Project
SM Smith et al., NeuroImage80, 144-68 (2013)
doi:10.1016/j.neuroimage.2013.05.039

You may also want to look through the contents of the special issue of NeuroImage80, 1-544, Mapping the Connectome, in which the above two papers were published. While I didn't use them for this blog post, the following two articles may be useful for context:

The WU-Minn Human Connectome Project: An overview
DC Van Essen et al., NeuroImage80, 62-79 (2013)
doi:10.1016/j.neuroimage.2013.05.041

Pushing spatial and temporal resolution for functional and diffusion MRI in the Human Connectome Project
K Ugurbil et al., NeuroImage 80, 80-104 (2013)
doi:10.1016/j.neuroimage.2013.05.012

The following papers provide some useful background on the use of blipped CAIPI for reducing artifacts in SMS-EPI:

Rapid brain MRI acquisition techniques at ultra-high fields.
Setsompop K, Feinberg DA, Polimeni JR, NMR Biomed. 2016 Feb 2.
doi: 10.1002/nbm.3478

Evaluation of slice accelerations using multiband echo planar imaging at 3 T.
Xu J, et al., NeuroImage83:991-1001 (2013).
doi: 10.1016/j.neuroimage.2013.07.055

A comparison of various MB factors with GRAPPA R=2 always enabled:

Evaluation of 2D multiband EPI imaging for high-resolution, whole-brain, task-based fMRI studies at 3T: Sensitivity and slice leakage artifacts.
Todd N, et al., NeuroImage. 2016 Jan 1;124(Pt A):32-42.
doi: 10.1016/j.neuroimage.2015.08.056