psychiatry

> I did not know that neurons had glycogen, but apparently most glycogen, at least, is within astrocytes and not neurons. That requires explaining why astrocitic glycogen causes a neuron's behaviour disorder identical to however bipolar disorder is "defined". Seems a tall order.

This requires a bit of unpacking. Roughly speaking, GSK3B deactivates glycogen synthase, which adds glucose to glycogen. That means inhibition of GSK3B would allow more glycogen synthase to function, adding more glucose to glycogen.

It's very interesting that you say that - astrocytes are implicated in bipolar disorder too. (See "Astrocytes in the Neuropathology of Bipolar Disorder: Review of Current Evidence", 2022, by Dai et al.) And astrocytes are a target of lithium ("Astrocytes are direct cellular targets of lithium treatment" by Rivera & [name I'll avoid], 2019).

But yes, a true causal link is an extremely tall order, and that's sorta my point. I'll elaborate below.

> "Computational biologists are trying to untangle the most complex “black boxes” on Earth.", I don't think they are, they are trying to imitate them, not understanding them. I think.

I strongly disagree, although I don't have evidence to disprove that (that said, you don't have evidence to disprove my original claim either). And, completely anecdotally, I know a few computational biologists and they're all trying to work on the "dissection" and not just imitating them. One scientist at my school has worked on papers finding genetic links between various diseases with the goal of finding links between their symptoms.

> I bet most of them know black boxes cannot be understood without dissection and looking at their innards directly, and experimentally with solid hypotheses grounded in scientific theory, not hunches and guesses from correlations and expert consensuses, starting from "constructs". But I might be wrong on some of that one.
> Looking for ideas in real science proceeds from having a scientific theory and unexplained observations that are falsifiable, objective and in principle amenable to experimentation, to look for ideas, why would anyone bother otherwise? ;)

Not necessarily, that's the point of abductive reasoning / "inference to the best explanation".
I'll continue my earlier analogy: we can see many of the items in the black box (e.g. GSK3B), but we can't see all of the connections (yet).
However, we CAN look at specific interactions within the box, e.g. "how does inhibition of GSK3B affect [protein X]?" That, IMO, doesn't require a solid hypothesis. At a certain point, it's easier to just do something. Why would a chemist hypothesize "what happens when I combine A and B?" As long as it'll be safe, why not just do it and see what happens? (That sort of "cause and effect" measurement is the sort of thing that bioassays, microarrays, RNA-Seq, organoids, etc. can be used for.)
And we can put those connections together to converge on the truth of components of the box. Think of it like trying to solve a jigsaw puzzle without looking at the box. Is it reasonable to predict what it looks like first? Can I predict the piece that's 4th from the top and 9th from the left? No, it's more sensible to just experiment and get a bunch of "chunks" of the puzzle. I can't put them together, but I know they DO go together.
This is reminiscent of a great scene from Futurama poking fun at "god of the gaps": https://youtu.be/UuIwthoLies?si=KIv-QJ1VmQc61X8n&t=73
I'm not mentioning that regarding to the religious point. My only point is that we can detect links between things even if we don't know the entire pathway. And we've only had the tools to look for those links for a decade or two.
If we take your points literally, most of the earth sciences aren't scientific either. I don't know of any way to find casuality of wind or ocean currents, so meteorology isn't scientific, right? But my weather forecast is still pretty good.

> "If you’re trying to solve a problem of this scale, you need as much data as possible.", not necessarily, more data can lead to spurious, non-causal correlations, that when incentives like publishing bias are present can lead fields astray for decades, or, knock, knock, centuries as in old medicine and psychiatry.

I'm not discounting biases - they're real problems. As another example of a bias: in the 60s (IIRC), physicists observed that new estimates of physical constants were very close to their "standard" well-known values. The new values should be more distant from the "standard" values because of errors in measurement. They do "blind analysis" for that. Or look at the debunked claims of superconductors.
Biases exist in any field, but throwing the baby out with the bathwater is a bad idea.
In fact, having the raw data of those experiments would be helpful - if it seems implausible (or if a scientist gets bored), they could rerun the analysis and realize something's wrong. That would be much easier than trying to discover bias through reading the papers. (And in the biomedical sciences there's been a push towards sharing raw data.)

> "massive alien computer[s]" do not exist ...

"Massive alien computer" was an analogy. But, in fairness to my point, the human body is a system that (at the chemical level) is "alien" to humans, and it's "massive" in the sense that all the hard drives on Earth combined couldn't contain a body. There are ~3x10^20 bytes of disk space on Earth. There are 7x10^27 molecules in the body (which would need to store, at minimum, 3 8-byte values each), so roughly 1.7*10^29 bytes.

> Making a lot of progress when at least half published research has, at least some, irreproducibility sounds disingenous to me. Sorry :)
Do you have proof of that in biomedicine? (Wouldn't surprise me at all if it's true of social psychology.)

> As for reading specifics about causality in mental disorders, it's not that I am dogmatic, but to me, aprioristically, the probability that ANY research paper related to psychiatry is correct decreases as more are published.

Let's say there are 2 papers. Do you mean that if I publish a 3rd paper, the odds that one of the 3 papers is correct is less than the odds that one of the 2 papers is correct? That's literally impossible.
So I'll assume you mean that the more papers there are, the less likelihood that any given paper is correct. I guess that's possible, but I don't have reason to believe that. Say there are 100 papers and 70(=70%) are correct. If I scaled that up to 1000 papers, why could I assume "aprioristically" that less than 700 are correct? There would obviously be more incorrect papers in absolute terms, but that doesn't necessarily mean a higher percentage of incorrect papers.

> Given the number of false ones has increased, and given the tendency without invoking ANY causality, agnostically that is, will most likely increase, that means if at somepoint in the future there is one that is actually true, it's likelihood of existing in the future, the ratio of that paper over all false ones approaches zero as time goes by... and publishing increases. Would you disagree stated thus? :)

Yes, I would strongly disagree. You're not only assuming that the ratio (new correct papers)/(new papers) decreases, you're assuming that it decreases to 0. As I said, I don't have reason to believe it'll decrease at all. Even I believed that, why would it decrease to 0? What reason to you have to believe it wouldn't "bottom out" at 50% (or 40%, or 20%) instead of 0%?

> So, if aprioristically there, out there, exists a research paper that proves beyond doubt ANYTHING about psychiatry it would be all over the news, I would know of it's existence without looking for it. Like my lost key Jedi-like falling into my pocket, without me recurring to Jedi mind tricks. :)

I hope, regardless of our political persuasions, we can agree that the media doesn't cover many, MANY topics that are noteworthy for some reason. When was the last time the 9-year-old civil war in Yemen got coverage?
And, in fairness, I understand why things like that wouldn't make the news - "nerds have an explanation for X" isn't very interesting because it doesn't directly affect anyone. There's a lot of developments in science (and other fields) that get drowned out because [some famous person] said/did [some silly thing].
"Nerds found an explanation for X and used that to make a medicine to treat it" is very interesting. That's why things like CRISPR treatments for sickle-cell anemia got at least a little press coverage.

>The only causality I think has been proven in psychiatry is the harm by medications, they cause "symptoms" similar or identical to the one's claimed to treat...
I can't find proof for that. And I have to point out - you seem very skeptical of science when it comes to links between molecules and psychiatric conditions, but you're accepting of a link between having PREVIOUSLY taken a molecule and the persistence of a psychiatric condition.

> people, particularly minors, might feel worse by being told they have a psychiatric disorder...
> there is one of teenagers embracing those labels without actually being diagnosed, like "fashion".
Don't these contradict each other? But yes, I definitely dislike kids claiming they're (X) without actually being diagnosed.

> Why would I invest time looking for it?...
> So, I will not devote time, short as it is ...
I sincerely understand. I don't know whether you've been diagnosed with anything, but many people won't look into something if they aren't affected by it. (And if they've already found a solution, they won't care either.) And I understand that perspective.

I'll share my own perspective briefly to add some context. I was diagnosed with bipolar disorder, Asperger's, and ADHD when I was young. I was harmed by a really bad doctor and medicine (she prescribed me Abilify and refused to take me off of a ridiculous dose). I changed doctors, I got on a new med but I was kept on a (lower) dose of Abilify. Things improved a bit for a while.

In 2017-ish, I changed doctors again and decided that I needed to deal with stuff myself. I started lifelogging in intense detail (quantifying EVERYTHING) and brute-forcing my way through research papers on my medicines myself. I've easily spent hundreds of hours by now. (I even guessed I have a mutation in a gene that metabolizes Abilify and some other meds. My doctor was sketpical, but I was right.) I finally got on a good medicine regimen in 2020ish and am doing much better. If I wasn't medicated or was just on Abilify, I would've put a toaster in my bath by now. But I'm an extreme aberration - a lot of people couldn't brute-force their way through those papers and just have to deal with bad doctors.
I research this stuff for fun - looking at the research I did to write this comment took an hour and a half. I even changed fields from computer science to computational biology because I enjoy researching this stuff. But I understand why people don't want to research it.
My concern is that some people only hear about the negative effects of medicines, so they assume that entire fields of research are bogus. If I only heard about the negative things about [X], I'd think it's terrible.

>But, those are eloquently put points. Thanks, I really enjoyed your comment.
Thanks for the compliment, and same to you, I enjoyed the conversation. Gotta call off this comment here after 2 hours.
I didn't see your reply until now (apparently I don't receive notifications). I'm find leaving this here, but if you want to discuss this further here's the base64 of my throwaway email:
eWFyaGFycjIxMg==
@[google's mail]