Distinct functional connectivity patterns in [ME/CFS & LC] during cognitive fatigue: a 7 Tesla task-fMRI study, 2026, Inderyas, Marshall-Gradisnik+

[SNT Gatchaman]

Distinct functional connectivity patterns in myalgic encephalomyelitis and long COVID patients during cognitive fatigue: a 7 Tesla task-fMRI study

Spoiler: Authors

Inderyas, Maira; Thapaliya, Kiran; Marshall-Gradisnik, Sonya; Barnden, Leighton

BACKGROUND
Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and long COVID are chronic debilitating illnesses featuring fatigue, post-exertional malaise (PEM) and neurocognitive deficits. Temporal correlation of neural activity between distinct brain regions, also referred to as functional connectivity (FC), can provide insights into how brain networks coordinate, at rest or during task. Therefore, we explored intrinsic FC correlates of cognitive fatigue in ME/CFS and long COVID patients during two Stroop-colour-word paradigms on 7 Tesla fMRI.

METHODS
450 sagittal volumes were acquired from seventy-eight participants: 32 patients with MECFS (pwME/CFS); 19 long COVID (pwLC) and 27 healthy controls (HC) during performance of baseline or Pre (before/during fatigue build-up) and repeat Post (fatigue set-in) Stroop tasks. Structural and functional data were analysed using the CONN toolbox.

RESULTS
Regions of interest (ROI-to-ROI) analysis revealed significantly increased FC in subcortical regions in HC for Pre vs Post. Relative to HC, pwLC showed significantly reduced FC between nucleus accumbens and vermis 3 (p = 0.02) in Pre and increased FC in the prefrontal cortex and hippocampus (p = 0.02) in Post. pwME/CFS showed a significantly increased FC between the left cuneiform nucleus and right medulla (p = 0.03). Compared to HC, reduced FC was significant in pwLC during Pre, and between medulla and hippocampus (p = 0.04) and between nucleus accumbens and vermis (p = 0.001) during Post. Aberrant FC was significant for pwME/CFS in core networks during Pre. Core network FC to the cerebellum, amygdala, caudate and red nucleus correlated with symptom scores for cognition in both pwME/CFS and pwLC. Hippocampus and cerebellar FC correlated with duration of illness in pwME/CFS.

CONCLUSIONS
Our findings of reduced dopaminergic hippocampal-nucleus-accumbens connectivity imply blunted motivation and cognition. Extensive FC differences in subcortical and core networks in patient cohorts were detected relative to an increased FC in HC. High regional communication indicative of greater task engagement by HC was distinctive while FC differences in ME/CFS and long COVID patients indicated reduced and dysregulated regional coordination that may serve as candidate biomarkers of symptomatology in long COVID and ME/CFS.

Web | DOI | PDF | Journal of Translational Medicine | Open Access

 

[DMissa]

Good to find that these results are now available for us to see. I am very curious to know what the people who understand these imaging methods think of this.

 

[SNT Gatchaman]

I've only skimmed intro and methods tonight. Imaging at 7T rather than the usual 3T could help with the low signal-to-noise ratio of fMRI. I think there may still be the problem of assuming a change in blood oxygen level equates to increased or decreased neuronal metabolism / activity, when it may be in the opposite direction in some regions. That may not be important if they're just looking to link areas which appear connected (or not) by temporally associated changes (regardless of up or down direction).

I see Wikipedia has a section on some of the potential pitfalls with fMRI.

However, deferring the technical stuff, I question the following —

the present study aimed to investigate the effects of cognitive exertion induced post-exertional malaise of the Stroop-colour-word task on differential FC patterns in pwLC, pwME/CFS and healthy controls

Two sessions of task fMRI were performed during a cognitively challenging Stroop paradigm. Session 1 of the fMRI task targeted at the initiation or buildup of cognitive fatigue and was referred to as the pre-interferenceinduced cognitive fatigue condition or Pre throughout the manuscript. Pre consisted of 60 Stroop stimuli, session 2 commenced 90 seconds after the end of session 1 when cognitive fatigue sets-in following Pre and was referred to as post-interference induced cognitive fatigue condition (Post).

Are they really trying to evaluate cognitive PEM after 90 seconds? It may not matter if it's actually evaluating rapid cognitive fatiguability (but shouldn't be termed "PEM").

 

[Trish]

CONCLUSIONS
Our findings of reduced dopaminergic hippocampal-nucleus-accumbens connectivity imply blunted motivation and cognition. Extensive FC differences in subcortical and core networks in patient cohorts were detected relative to an increased FC in HC. High regional communication indicative of greater task engagement by HC was distinctive while FC differences in ME/CFS and long COVID patients indicated reduced and dysregulated regional coordination that may serve as candidate biomarkers of symptomatology in long COVID and ME/CFS.

Does the use of terminology I have bolded here worry anyone else?

To me it raises the unhelpful and damaging term 'effort preference' from the NIH Walitt study. In both this paper and the Walitt one, the observations don't warrant the implying of voluntary lack of effort. Surely the brain function changes they describe from the fMRI are physiological, not psychological.

I couldn't see any data on how the people performed in the Stroop tests. Did I miss it?

I also noticed the misuse of the term PEM. They are observing the effects of cognitive fatigability, not PEM.

A lot of the p values are quite close to 0.05, and I can't see any mention of correcting for multiple comparisons.

 

[Eleanor]

Does the use of terminology I have bolded here worry anyone else?

To me it raises the unhelpful and damaging term 'effort preference' from the NIH Walitt study. In both this paper and the Walitt one, the observations don't warrant the implying of voluntary lack of effort. Surely the brain function changes they describe from the fMRI are physiological, not psychological.

Yes, and:

Negative association between these regions in our findings substantiate the presence of apathy in pwLC, i.e. a lack of interest and enthusiasm towards solving the energy-demanding cognitive exercise.

Funny how these apathetic and uninterested people managed to sign up for and participate in the study, completely voluntarily...

Also the speculation about 'response to trauma' here is weird.

increased connectivity correlations between the two regions in pwLC during the Post-interference condition may point towards an adaptive
response to the complexity of the Stroop task. The act of recalling and imagining the future would also most likely result in increased communication between the two regions [111, 112]. These patterns were observed in a post-traumatic stress disorder (PTSD) study where patients with PTSD showed increased hippocampal-mPFC connectivity as an adaptive response to trauma [113].

 

[Sean]

Negative association between these regions in our findings substantiate the presence of apathy in pwLC, i.e. a lack of interest and enthusiasm towards solving the energy-demanding cognitive exercise.​

Which is exactly what you would expect to see in people who cannot do those task due to acquired pathophysiological limitations.

Indeed, it is an unavoidable behavioural adaptation. Quite literally all they can do.

 

[Trish]

Apathy is a pejorative term to use here. If they were truly apathetic about the task, they wouldn't have volunteered to take part in the research.

What about exhaustion, inability to focus eyes and brain on the task, slowed reaction times, feeling very unwell and in pain. Difficulty making sense of the image on the screen?

I find myself lying here staring at the screen sometimes for seconds, sometimes for minutes, while writing posts or doing mod tasks or reading documents increasingly frequently and for longer periods the longer I try to concentrate. It has nothing to do with apathy, it's a slowing or stopping of ability to keep functioning.

 

[Sean]

To continue attempting a task, however imperfectly, and at high effort cost, requires high motivation and intention.

And to decline to do so may well be the only rational choice.

But no, the authors have to use prejudicial words like apathy, and lack of interest and enthusiasm.

Really, they couldn't come up with some more neutral words?

 

[Eleanor]

To continue attempting a task, however imperfectly, and at high effort cost, requires high motivation and intention.

And to decline to do so may well be the only rational choice.

But if the brain activity necessary for the task isn't happening, or is impeded from happening, then choice doesn't enter into it (rational or otherwise). it's just not happening.

 

[Eleanor]

The vermis is a region with diverse cognitive, motor and emotional processing roles [106, 107]. Reduced FC in this region, as observed in pwLC may possibly be due to increased apathy and reduced engagement during the Stroop task.

is there any legitimate reason to assume the direction of causality here?

 

[Kitty]

Does the use of terminology I have bolded here worry anyone else?

Yes, and the findings surprise me a bit. I'd have predicted seeing people with ME/CFS using more of their total cognitive resource on the task because every process is so inefficient.

I know brains don't work mechanistically, but there is somehow a cognitive equivalent of physical momentum. Starting a movement needs more energy than sustaining it, so if you're forced to keep stopping, the total fuel required to move from A to B will be greater. If you can't maintain focus and have to keep regathering it, the resources consumed are greater.

I keep thinking we should rename brain fog as brain entropy.

 

[V.R.T.]

This is all very Wallitt-esque. Disappointing.

 

[bobbler]

Negative association between these regions in our findings substantiate the presence of apathy in pwLC, i.e. a lack of interest and enthusiasm towards solving the energy-demanding cognitive exercise.​

Which is exactly what you would expect to see in people who cannot do those task due to acquired pathophysiological limitations.

Indeed, it is an unavoidable behavioural adaptation. Quite literally all they can do.

And who find that noise gives them migraine , pain, exhaustion, cognitive issues and are in a noisy MRi on top of the executive function load of being in pain from lying down and dealing with other symptoms whilst doing any tasks. So it isn’t like with like. No HC with these issues that would make an mri tough would sign up for it.

 

[bobbler]

Does the use of terminology I have bolded here worry anyone else?

To me it raises the unhelpful and damaging term 'effort preference' from the NIH Walitt study. In both this paper and the Walitt one, the observations don't warrant the implying of voluntary lack of effort. Surely the brain function changes they describe from the fMRI are physiological, not psychological.

I couldn't see any data on how the people performed in the Stroop tests. Did I miss it?

I also noticed the misuse of the term PEM. They are observing the effects of cognitive fatigability, not PEM.

A lot of the p values are quite close to 0.05, and I can't see any mention of correcting for multiple comparisons.

If there is dopamine differences it is linked to far more and I’ve always thought the assertion of ‘motivation’ was simplistic and perhaps the wrong nuance for what ‘that thing they see’ actually is.

I suspect (given the upsurge post-workout) it’s related to ‘flow’ of enjoying the activity/getting satisfaction etc

Which makes far more sense when you think about the pain from noise and exhaustion for pwme vs hC etc

 

[jnmaciuch]

A lot of the p values are quite close to 0.05, and I can't see any mention of correcting for multiple comparisons.

In the methods it mentions some kind of special FDR procedure specific to this type of imaging.

What I’m still very unclear on even after going down a rabbit hole through references and analysis walkthrough is which tests were corrected together. Maybe that’s just because I’m very unfamiliar with the field.

They list 67 regions of interest, and then apparently the data points are pairwise combinations of regions. I am not sure if a voxel/cluster means the same thing as a region of interest here. Then there are comparisons between pre and post timepoints within each group, and between group comparisons at both timepoints (9 total). The methods state that FDR correction was done at the cluster level. What would be crucial to know is if this “cluster-level” correction was across all paired-region datapoints for 9 comparisons or if things were treated as separate hypothesis. If someone knows how these analyses are typically done I would really appreciate some insight.

 

[SNT Gatchaman]

On the cohorts, the healthy controls are up to a decade younger. One of the ME/CFS patients (#56) was 55.6 with illness duration of 48 years. Most of the LC patients' illness duration was under 2 years (as might be expected) but many are under 1 year and a few 3 months or less —

LC
Age 46.20 ± 12.62
Duration 0.8 ± 0.6 yrs

ME/CFS
Age 42.9 ±10.5
Duration 12.7 ±11.0 yrs

HC
Age 34.75 ± 8.39
Duration N/A

A little bit more on technical limitations with this technique. They have a page on denoising and quality assessment of denoising. After this process 10 outliers were excluded: 3 (11.1%) HC; 2 (10.5%) LC; 5 (15.6%) ME/CFS.

I'm not sure these techniques have sufficient spatial and temporal precision. Or reliability. I would expect that replication attempts with different cohorts would show quite different regions with hyperconnectivity and hypoconnectivity. I guess one thing to do would be to compare with other studies of healthy controls doing a similar cognitively fatiguing task.

It was probably reasonable to conclude with —

Viewed in aggregate, the evidence suggests that functional impairments and cognitive dysfunction may arise from reduced energy metabolism and present as executive and memory problems.

But going on to —

Both long COVID and ME/CFS patients showed impaired functional connectivity among regions of the cerebellum, vermis, and deep grey matter during interference-induced cognitive fatigue which is indicative of cognitive exhaustion. This suggests reduced motivation and a decreased dopaminergic influence thereby inducing neural exhaustion.

seems to be drawing a pretty long bow. Why the need to invoke motivation? Surely reduced energy metabolism alone would be sufficient to induce neural exhaustion. What, hypothetically, would this test look like in someone with diabetes and acute hypoglycaemia? And for all we know those BOLD signals may in fact be showing the opposite effect in terms of regional metabolism, that it's actually increased in some regions in the patients attempting to compensate for impairments more generally.

I mean it's pretty obvious looking at the daily discussions on S4ME over a decade that one thing ME/CFS patients absolutely do not lack is motivation (and interest and desire and...).

 

[bobbler]

On the cohorts, the healthy controls are up to a decade younger. One of the ME/CFS patients (#56) was 55.6 with illness duration of 48 years. Most of the LC patients' illness duration was under 2 years (as might be expected) but many are under 1 year and a few 3 months or less —

LC
Age 46.20 ± 12.62
Duration 0.8 ± 0.6 yrs

ME/CFS
Age 42.9 ±10.5
Duration 12.7 ±11.0 yrs

HC
Age 34.75 ± 8.39
Duration N/A

A little bit more on technical limitations with this technique. They have a page on denoising and quality assessment of denoising. After this process 10 outliers were excluded: 3 (11.1%) HC; 2 (10.5%) LC; 5 (15.6%) ME/CFS.

I'm not sure these techniques have sufficient spatial and temporal precision. Or reliability. I would expect that replication attempts with different cohorts would show quite different regions with hyperconnectivity and hypoconnectivity. I guess one thing to do would be to compare with other studies of healthy controls doing a similar cognitively fatiguing task.

It was probably reasonable to conclude with —



But going on to —



seems to be drawing a pretty long bow. Why the need to invoke motivation? Surely reduced energy metabolism alone would be sufficient to induce neural exhaustion. What, hypothetically, would this test look like in someone with diabetes and acute hypoglycaemia? And for all we know those BOLD signals may in fact be showing the opposite effect in terms of regional metabolism, that it's actually increased in some regions in the patients attempting to compensate for impairments more generally.

I mean it's pretty obvious looking at the daily discussions on S4ME over a decade that one thing ME/CFS patients absolutely do not lack is motivation (and interest and desire and...).

For that matter it would be interesting if the odd research used as controls people who are as exhausted as I can even begin to describe

That metaphor I use being for them to imagine having flu or tinsilitis or something genuinely bad with glands up and aches then being stuck in transit for 5days in economy with only airport lounges and plane seats to sleep in or ‘get rest’. It doesn’t touch it of course because you’d also have to add in taxing work mentally and ensure there were noises that bore into the head probably but that’s your starter

For when I try and get people to realise I’m not thick or ‘unable to think’ it just hurts and is slow and if is being demanded of me unnecessarily should one day be termed abuse … because I’m struggling for finding words and maybe even stuttering if they are having to continue with ever more complex explanations someone demands just as they would if they’d got home needing to collapse after having had a really really mad week of 20hr work days where it was full on questions multi tasking from all angles.

I know we can’t give people flu but I don’t understand how they can have the gall to rudely assert motivation when they aren’t comparing situations to situation . At the moment it’s like as if healthy controls being used to show up controls who’ve just done one of the above for the last week, so we’re exhausted (like pwme are getting to an appointment) as ‘demotivated on the same task’ - they aren’t studying people who’ve done the same thing due to the exertion of getting there and lying there in discomfort that HCs don’t have.

So I think genuinely that if they are using comparison of which areas light up etc and should be to do scientific psychology then be removing such obvious variables that should be controlled for then they need to run that experiment so they can see what non ill people in exhaustion and maybe even pain whilst doing it look like on the scan and only then subtracting to see if there is any differences which could be said to be down to ‘the people’ or ‘the illness’ in the illness group.

 

[DHagen]

Perhaps I am being too generous here, and I am certainly speaking from a position of ignorance, but looking over other neuroscience papers dealing with "blunted motivation" and "motivation" more generally (e.g. with regard to Parkinson's), it seems possible that "motivation" is here to be understood as essentially synonymous with the neurological mechanisms by which impulse or need is translated into action, removed from the idea of "will" or "desire" as employed in common speech. In this way, the argument would not be "these people don't really want to do the activity, so they're bad at it," but rather, "the mechanisms by which activity is engaged and sustained via dopaminergic connectivity are not functioning properly."

Or perhaps it is as inflammatory as it seems and the connection between dopamine and "motivation" is simply so ingrained that any study identifying a problem with dopamine is going to be read as essentially a problem with "motivation."

Regardless, if the authors' speculations can be set aside, I am still quite interested in some of the findings here. The >10 year age difference between HCs and the LC/ME cohorts does seem like it could be a pretty significant issue, however, particularly given that (as I understand it) aging has some pretty significant effects upon dopamine activity.

 

[Hutan]

This paper is certainly puzzling. Leighton Barnden has generally been regarded as a decent investigator in the field of brain imaging. Some of his previous work in ME/CFS has been focussed on the brainstem. The work of Sonya Marshall-Gradisnik (co-leader of NCNED) has been criticised on the forum, but she and the rest of NCNED have mostly been solidly anti-BPS e.g. anti-GET.

So, I'm not sure where the impetus for using judgemental words like apathy have come from. And, at the very least, this team should have understood what PEM is.

What I've read of the paper hasn't shifted my usual skepticism about these sorts of studies. As others have noted, there was a very large number of potential comparisons, the p values are not flash and the cohorts are not well matched. There doesn't seem to be much in common between the LC vs HC and the ME/CFS vs HC results, and I think there should have been if there were real results here.

So, yes, I'm a whole lot less interested in the reported findings of the paper than I am in how this team came to put the spin on the results that they did.

 

[DHagen]

There doesn't seem to be much in common between the LC vs HC and the ME/CFS vs HC results, and I think there should have been if there were real results here.

Perhaps this goes too far in piling one hypothetical on top of another, but if there was something real being measured here, could the disparate results be in some way connected to the "opposite white matter abnormalities" that Yu et al thought they found in comparing pwME and gradual onset with pwME and acute onset (assuming that all of the pwLC would fall into the latter category)?