Grip test results and brain imaging in the NIH study: Deep phenotyping of PI-ME/CFS, 2024, Walitt et al

I've posted some relevant papers under the electromyography and dimitrov tags. (Edit: @Hutan including alternative links where necessary)

The figure shows initial, last non-fatigued and the first three fatigued blocks. It would be interesting to see the entire series (the raw data for 4A only includes these same blocks), but to my eye that looks like a smooth increase in the decline with time — the sort of thing you might see if there was a physiological fatiguing process occurring, rather than some sort of effort preference decision.

But we'll have to read the papers to see how these conclusions are made.

 

This is very important for interpreting results. It's a tiny study. I don't think a meaningful comparison of HV vs patients is possible for a complex pattern like the Dimitrov Index. I suspect any conclusions based on the difference are not statistically valid.

Correction: the figure says it is statistically significant:
e; n = 6 independent participants) and PI-ME/CFS (red; n = 8 independent participants) patients. A significant difference was noted between the groups (0.2 ± 0.5 versus −0.43 ± 0.3, t(12) = 3.2, p = 0.008)
I'd be interested to hear how you would statistically test to see if the pattern is different when it fluctuates wildly in both groups
Also, the foundational dimitrov index study was based on 7 subjects (leg muscle) and noted
"Further studies are needed to fully establish the validity and reliability of the new spectral indices for testing muscle performance in the clinical, rehabilitation, and sports setting."

Added: as @Hutan said, the pattern is similar for both groups - given the tiny sample size, this looks like the same basic thing is going on for both.

 

Copied post

This is nonsense but as a necessary upbeat take on a very costly study that was supposed to be groundbreaking it is excusable, and at least a good message to the world.

This I think is more problematic, not only nonsense but unhelpful. As far as I can see they have simply shown a weak correlation between some task choices and some brain signals. Those task choices may well be entirely healthy in that they may reflect a sensible learnt pacing strategy. Or maybe just rational choices. It is far from clear that the control choices were more rational.

Moreover, the brain signals simply suggest that the patients were thinking differently. Maybe they were thinking less about which bar to go to later or more about how many changes of clothes they had brought. Nobody has any detailed idea what these signals mean. A reasonable generous interpretation that allows that the signals might actually have something to do with the choices is that they show healthy rational thinking as opposed to whimsical thoughts in controls for whom none of it matters.

It is possible that the signal patterns are a tell tale sign of whatever is actually stopping PWME from doing things but if so nothing has been discovered about what that is or where it is.

 

Definitely or as others have mentioned it might not really be relevant at all. Perhaps having to worry about the real threat of PEM causes that part of the brain to act in a different way. Maybe they are already experiencing the effects of PEM and the pain/discomfort/and associated symptoms with this causes some fmri changes.

If this is the case, it is not broken and also not doing anything to influence activity levels. I personally believe this doesn't seem to be particularly important to the pathology of ME/CFS at all.

 

This is another one I'd love to pick into as you've said, and it all involves peeling out so much context given what they were putting these pwme through (I'm assuming most people agree that I doubt there were many points where the pwme were not in some sort of rolling PEM, along with all the other phenomena we put up with like sensory and being in an uncomfortable bed and so on layering onto each other). So it's context on context.

It feels really weird when you imagine the potential executive function load/'functions' going on of someone trying to 'stay upright' (and even that must have been in a metaphorical sense for some) whilst adding in what investigators perhaps thought were just the tasks set, but as we know would have been complicated by vertigo, nausea, pain and discomfort and people thinking about whether they need the loo vs when it will finish and all the planning involved with that and so on.

I think the reason that I note this, as someone who is not an expert at all is whether there are phenomena that either are or aren't well studied at all where the brain has to effectively do what parts of the body might do in certain situations where it eventually has to 'prioritise towards the most life-sustaining type things' or is simply dividing thinly and having to go in order. Which is quite different to eg hibernation type theories and more a version of overload due to how many things are going on due to what the body, already ill, is being put under. EDIT: as simple as 'you can't physically direct someone to 'pay attention to/focus 25 things at once when they are all in different places and directions'' just simply because ti is a contradiction in terms.

That of course wouldn't necessarily be an issue 'with or caused by the brain' but I guess something you might call downstream but is more about not acknowledging the additional load going on, and if you'd induced sea-sickness and had the HVs awake for 48hrs before before putting them in a chair that holds their back at an uncomfortable angle then you might have seen similar things.

 

On the surface this sounds like evidence of peripheral fatigue, right? The participants are told to try to squeeze as hard as they can for as long as they can but they can't do it for as long, I would assume because their muscles became fatigued more quickly with time. This also happens to the healthy controls, it's not like they can squeeze forever. Can someone help be understand what evidence they cite that proves to them that the muscles weren't just becoming fatigued faster?

 

I think it's this:

So they think that pwME/their brains (depending on which author you ask) choose to not maintain grip force despite being able to. If you think of this not as a choice, but as a kind of brake in the brain that is being activated for reasons unknown (but perhaps immunological or neurological), it is interesting. That brake presents to pwME as escalating symptoms. When it's presented as "Aha! It's not that they can't, it's that they won't!", without a good explanation of what's happening, it's odious.

I think the PEM after CPET results show that that brake is there for good reason - when pwME push through the brake, ie do what they "can" as opposed to what they "will" because people provided "strong encouragement" - they get this:


from the separate PEM after CPET paper from the same study:
Stussman B, Calco B, Norato G, et al. Mixed methods system for the assessment of post-exertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome:an exploratory study. BMJ Neurology Open 2024;6:e000529. doi:10.1136/bmjno-2023-000529

 

And just to qualify all of this - if the effort task does not show what they think it shows, ie effort preference, then this nice story falls apart. If the effort task is detecting a real difference between pwME and healthies, then this should show up in bigger samples with the task modified for individual calibration, fixed number of trials, fixed duration of trials - basically a task where

 

Hey everyone, new to this forum but I have lots of questions with the way this study was performed/interpreted. I'm glad others have noticed the disparity between the analysis of the fMRI data versus the MEP data, specifically by not standardizing the comparisons across fatigue onset. I've been approved for access to the data repository so I am going to attempt to re-analyze it to see if there is still a significant difference in TPJ BOLD signal when this is controlled for, as the most significant difference seems to be during two sets of blocks where the majority of HVs had not reached muscle fatigue while most of the PI-ME participants had.

I am also incredibly puzzled by the interpretation of the TPJ results. The paper that Walitt keeps citing for this is a recent hypothesis paper, and reading it has left me with far more questions than answers. I'm a neuroscientist but I mostly deal with the realms of pharmacology, cell biology, genetics, and animal behavior; cognitive neuroscience and neuroimaging have always been a bit too theoretical for my hyper-literal brain to fully grasp. So I'm gathering as much as I can but hoping to connect with someone who works in this area and can help explain the logic behind the interpretation of the data.

So first of all, the references: ref 31 is the hypothesis paper about the functionality of the TPJ, and ref 32 is a research study from 2011 that tested self-agency by having participants wear a specialized glove that can basically transmit the data on its position and display a hand figure on a screen that the participant watches while undergoing the fMRI. They would then reduce the accuracy of the figure so that the proprioceptive information (where your brain thinks your hand is in space) and the visual representation (the figure on the screen) are in conflict. The TPJ seems to be one of them, but there are bunch of others in this network (they mention PPC, STS, DLPFC, pre-SMA, precuneus, insula, and cerebellum).

Basically, from this we've learned that a mismatch between proprioception and visual feedback can activate the TPJ. The hypothesis paper discusses a lot of studies that found TPJ activation via fMRI in certain contexts like "theory of mind" studies (basically the ability to understand someone else's state of mind based on context), episodic memory tasks, and attention/orienting tasks like the Posner task. But then this quote stuck out to me, emphasis added:

I'm interpreting this to mean that rTPJ is activated when sensory feedback does not align with expected performance, which results in feeling a lack of agency. Importantly, this loss of agency has been induced either by manipulating visual feedback (i.e. using virtual reality) or by directly stimulating/deactivating the TPJ via TMS. I did a few PubMed/Google Scholar searches to see if any studies had looked at TPJ activation during repetitive motor tasks to try and determine whether reaching a point of muscle fatigue can also induce the perception of a loss of agency. To my mind at least, it seems like there could be a difference between a proprioceptive/visual mismatch and the failure to perform a motor task due to muscle fatigue. I didn't find anything - that doesn't mean it doesn't exist, but it does mean that the conclusions Walitt et al draw from this result are basically speculative, right? This wouldn't bother me if it was framed as such, but the paper discusses it as though it's widely implicated during muscle fatigue, which it does not appear to be in any way.

(I'll continue in another post, getting tired and this is getting long but I have many more thoughts I'd like to share)

 

@Janna Moen PhD fantastic that you have access to the data. If you also have institutional access to a neuro(vascular)radiologist with good experience in BOLD fMRI, ask them about —

 

What about a shift in how certain brain cells are utilizing energy. If there is some subtle shift in the way energy is generated perhaps cells wouldn't demand as much oxygen to replenish their supplies. This could also lead to a decreased BOLD fMRI signal.

 

Yes, neurovascular coupling is always an important caveat in BOLD fMRI studies, often an unappreciated one. There were a few labs studying this where I went to grad school. It's really difficult to measure experimentally and is basically impossible to measure accurately via fMRI alone - the lab I know about was studying coupling in mice using some incredibly fancy laser doppler scanner with a two-photon microscope to measure the way blood vessels responded to neural activity. It's one of those baked-in assumptions in almost every fMRI study (that neuron activity = increased blood flow) and you will get weird looks and shrugs when you point out it can also be a confounder. I would love to talk to Michael VanElzakker about this as he has much more expertise in this area. Another caveat is that fMRI sampling is fairly limited temporally, you can only acquire new image stacks every ~2s, so any changes that might occur under a faster timeframe (i.e. milliseconds) will be unmeasurable.

 

I’m curious if one of the values @andrewkq has evaluated more strongly correlates with this.

Agree.

Gender, age, etc. are not given for these in raw data.

And even in the first grip test, even if the groups are well matched, the variance due to gender, age differences within the groups weaken the p-value. Ie they can show no difference when there is difference.

Known factors should be modeled and controlled for.

 

For the full sample, time to failure correlates with hard task completion rate for both the dominant and non-dominant hand, but the latter should be more relevant since that's the hand used for the hard task. Caveat is that Kendall's tau is barely not significant for non-dominant. Both dominant and nondominant were non-significant when looking at only ME or only controls, which is how Walitt reports the findings for PHTC. Here are the numbers for the full sample.

Dominant Hand Time to Failure vs Hard Task Completion Rate

Log Transformed Pearson: r(26) = -.50, p = .0074
Spearman: r(26) = .52, p = .0048
Kendall: r(26) = .39, p = .0073

Non-Dominant Hand Time to Failure vs Hard Task Completion Rate

Log Transformed Pearson: r(26) = -.40, p = .033
Spearman: r(26) = .40, p = .035
Kendall: r(26) = .28, p = .059

 

Has someone else mentioned the obvious of the gender differences in the groups? And whether they varied the size of what was being ‘gripped’ to hand size given men’s hands are generally larger. It’s pretty hard to exert the same pressure when you can barely get your hand around an instrument and I’m imagining something like a petrol pump, where I have to use both hands initially to get the lever part ‘in’ to the point it can be held by one hand as my hand can’t stretch enough to even reach the other bit with all fingers as it is.

my hand also wouldn’t ‘close’ round the whole thing like others around me do so the task is totally different when you are trying to use the end of your fingers to keep the lever in place vs the force of it springing back out