Use of EEfRT in the NIH study: Deep phenotyping of PI-ME/CFS, 2024, Walitt et al

thanks @Murph & @andrewkq

that makes sense

 

Exactly! And I think this must be the article I tried to refer to in my just-now post on the old Dan Neuffer book, the paper my MIL read that was mentioned in a Health Rising newsletter.

 

I think that is exactly right.

It is quite hard to find a clear way of expressing one's intuitive sense of 'no way' about this but that is a good one.

Another way maybe is to point out that the authors are wrong to talk of the brain deciding things. The brain is a forum within which information is passed from one part to other parts all the time. A decision on a choice will involve some component of the brain choosing but also another component of the brain in the same place presenting the chooser with the odds of better or worse to make the choice on.

An abnormality of a chooser component is a completely different thing from an abnormality of an odds presenter component - which may tell the PWME's chooser component that the pay back is going to be bad. But there is no way you can tell from brain scans which component is wrong because there is only one place where one is providing information to the other.

The fundamental problem here is that the psychologists have no understanding of the structure of the mind. My other main interest is in just that - the relational structure of thinking processes. The authors make the standard false assumptions.

 

Agree with everything you said. Healthy volunteers have only rewards. Patients have both rewards and penalties to consider. Even if you argue that healthy volunteers could have non-dominant arm/hand pain/stiffness the next day if they do lots of hard tasks, they also have the experience of healthy people that this will not stop them doing what they want, and will resolve nice and quickly. Whereas patients are reckoning with days/weeks of consequences both from this particular task viewed on its own, and from the cumulative effect of a week of repeated overexertion, and another to finish the study. Edit to add: the consequences of participation in the entire study could go on for months/years for some participants.

The reason this is not the central part of the discussion here is, in my view, that we don’t have relevant data to draw on to make the argument stronger, because PEM wasn’t measured in the effort task.

The strongest support for your argument comes from their own study of PEM after CPET. Patients reported climbing malaise after CPET, peaking 24 hours after baseline and still leaving them “much worse than baseline” at 72 hours. They didn’t measure beyond 72 hours which is a shame. We should also point to other studies that have reported PEM after tasks, ideally in submaximal tasks ie those that involve effort levels a bit more similar to the effort task.

We can point to that, and say that the EEfRT task, and specifically hard tasks within EEfRT, are likely to have been difficult enough for some pwME that they would have expected it to induce PEM, and thus that this potential penalty would have factored into their choices between hard and easy tasks. From a scientific point of view, it’s only a hypothesis – the hypothesis that patients were factoring in penalties that healthy volunteers weren’t. And this may not seem plausible to many scientists: people without PEM would find it easier to imagine PEM following CPET than PEM following repetitive finger movements. Since we don’t know what the participants were thinking, and it hasn’t been reported, the argument is left on weak ground.

Given that, the strongest way to refute their finding would be to demonstrate, with their data (or someone else’s), that something else, ie something other than pwME “avoiding” hard tasks, explains their finding better.

But the reward-penalty argument you are highlighting absolutely has to be part of any response. I would argue that it should be the first argument made.

Edited to add: the consequences of participation in the entire study could go on for months/years for some participants.

 

I see that point entirely.

But it also works to invalidates the thing, because if they're not measuring PEM they're not measuring ME.

 

I think that's why they're not attributing patients' performance to anhedonia. They're saying, it's not that patients aren't reward-driven, it's that they're driven to avoid effort.

 

The point of the EEfRT is to somehow magically capture intrinisic psychological motivation phenomena, not strategies.

Technically/Strategically the probability that is assigned to a task should not matter at all for a healthy person (assuming finger fatigue doesn't crucially matter). It should be irrelevant to them whether a task with $4.00 has a probability of 0.12 or 0.88 because in either case it's a "free lunch" and since they don't have an energy limiting condition it really shouldn't matter if they spend energy on a task even if the probability is low, it's always a "free lunch" no matter what the probability is, what matters is the size of the reward. An example from everyday life would be given that chance to play in the lottery for free, anybody (that is keen on winning some money) sensible would have to say yes.

But given the psychology of the game and that not everybody will understand it in it's full depth, there's still a trade off with people saying no to a low probability task even though there is no intrinisic reason to do so, so it does capture some psychological and or biological phenomena, whatever these may be. I think the GEE captures this better by analysing all interactions than your analysis.

For someone with ME/CFS the situation may of course be entirely different, as they may have to manage their energy and as such "free lunches" stop existing.

That's why the authors state that the patients aren't driven by potential reward, rather than that they are driven by a general lack of effort, so to speak in Walitts analysis. The EEfRT is supposed to remove strategies and supposed to capture some intrinsic nature of people, the nature the authors argue was captured here is that pwME prefer to go hard a bit less often.

 

Yes, that's true. Haha, I had actually thought about exactly that before against the counterargument I proposed above when I thought about it for the first time a couple of days ago, but then forgot about it again. You are of course right. I will edit my comment above.

 

Or even better, do the quick easy task and fail it so it doesn't go into the pool of rewards

 

Something I did want to re-mention about these findings an the reward drive which is also what your analysis is about and which is something I wrote about a very early comment here and which is something others have also talked about is that in a EEfRT study you measure a tremendous amount of things so it's very likely you'll find one thing that is statistically significant, but then can't be reproduced in further studies (which is exactly what happens in a lot of studies) and that you can put any twist on the results that you want to.

That is why the conclusion of these results could have been precisely the opposite as well and the kind of analysis you, myself and other people have done on here has been used to do exactly that and there's precidence for it in other studies "Thus, individuals with schizophrenia displayed inefficient effort allocation for trials in which it would be most advantageous to put forth more effort, as well as trials when it would appear strategic to conserve effort.", i.e. it can be just as well be argued that going for hard more often, but without it being in the "right moment" shows an inefficiency effort allocation, i.e. if one could show this via a rigorous analysis one could show that pwME have a higher functioning effort allocation rather than anything else.

 

Yes, I thought about that too and proposed precisely that in this comment:

However, when I began looking into it nothing too fruitful came off it, at least at a first glance with a basic analysis. The variance amongst the pwME is quite high so you also have pwME choosing to go hard-hard-hard-hard-easy-easy-easy-easy-hard-hard-hard which destroys more obvious general patterns which would be indicative of "energy conservation" or "breaks". Maybe a less basic analysis could reveal something, but I really have no idea and also don't know how strong of an argument it'll end up being rather than being in support of "effort preference".

 

Definitely a possibility. In the repetitive grip testing by functional imaging they report:

Maybe reduced blood flow to the brain could be what causes pwME to stop being able to exert?
Fwiw, I had cerebral blood flow measured during a tilt test and it was supposedly fine. (I fainted quite spectacularly, though only at 32 mins.) The sensors were only place on the left side of my head, though - I don't know if they would have picked up reduced flow on either side or only on the left side. I was disappointed as I get all kinds of weird sensations on the right side of my head - tingling etc - and absolutely none on the left. My understanding is that there are a number of different ways to measure cerebral blood flow. I have no idea if the method used would be considered a good one or not.

Agree.

The closest I can find in the paper is this:

In my view that describes a voluntary behaviour. But as lots of us have been pointing out, they're thinking of fatigue only as something that isn't there at the beginning of a task and might develop during it, not as what pwME come into each task with.

Agree, there must be a reason why some studies started doing individual calibration. @andrewkq , am interested to know if you have any thoughts on this (which I wrote above in reply to @EndME 's good point):

Thank you so much for explaining all of that. Much appreciated.

Agree. Would be nice if some of the literature touched on these points, I haven't been able to delve in yet...

Yep, possible. Based on Walitt's pre-existing views on CFS, and the fear-avoidance/deconditioning model, I'd say this was pre-hoc.

Exactly.

Exactly.

Agree.

Great find. I think it's very relevant - any evidence that pwME could have been expected to enter this task at a motoric disadvantage compared to the healthy volunteers supports the idea that the task as operationalised by Walitt et al could not assess preference without individual calibration.

 

Have you been fraternising with HV F?

 

Guess that doesn't matter, if you can get more grants to try.

 

Evergreen thank you sooo much for your detailed response.

This makes to sense to me, you cant prove one hypothesis wrong with another hypothesis, no matter how obviously correct it might seem to those who have lived experience of the thing occurring. It is too easy to break down.

You cant disprove A by presenting B which is also unproven.

I am neither physically or emotionally robust enough to write letters, nor do i have the professional scientific language to do so, but i am so grateful for all the work that is going into all this analysis, i cant understand most of it as am too cognitively impaired at moment, all the maths and charts etc are completely opaque to me. But i am massively grateful that others are doing it, and i sacrificing their time & energy to do so.

Thank you all for hearing me and explaining, glad to know i not way of the mark