I’m mildly skeptical that blindness prevents schizophrenia

[-]AnthonyC3y152

Is there anything about the way schizophrenia is (or used to be) diagnosed that would make it harder for the congenitally blind to get diagnosed? I ask because I know someone, completely deaf from birth (and who only learned sign language as an adult, not sure if that makes a relevant difference in terms of language processing), who for a long time couldn't get treatment for (and never got a formal diagnosis of) schizophrenia on account of a lack of auditory hallucinations or hearing voices.

[-]Steven Byrnes3y30

Thanks! Yeah, I’m not an expert, but if God / Omega told me with absolute certainty that there were somewhat fewer documented congenitally-blind schizophrenics than expected, I would definitely start brainstorming explanations like “Maybe schizophrenia presents a bit differently in congenitally-blind people, making it hard to diagnose?” or “Maybe blind schizophrenics are less likely to wind up seeing a psychiatrist for some reason?” or things like that.

I don’t think those kinds of things would amount to an orders-of-magnitude reduction in documented congenitally-blind schizophrenics. But if we’re trying to explain a 50% reduction or whatever, sure, seems possible.

The thing is, I don’t think anybody is claiming that we’re trying to explain a 50% reduction (?). These people seem to be saying that there are orders of magnitude fewer congenitally-blind schizophrenics than expected. Whereas my tentative read of the evidence is: “There could be anything from 10× fewer to 3× more congenitally-blind schizophrenics than expected from chance.”, i.e. there might not be any reduction to explain in the first place.

[-]Scott Alexander3y140

Thanks, this had always kind of bothered me, and it's good to see someone put work into thinking about it.

[-]ChristianKl3y60

Well, so much for that plan!

An incidence rate of 1 in 2 million might not be enough for Danish (5.831 million) and Western Australian (2.667 million) but might be enough for the United Kingdom (67.733 million).

The United Kingdom does want its data to be analyzed. I know an EA who works as a contractor to do ML to help them with data analysis. I however don't know the exact rules under which data analysis happens. Message me if you want the contact.

[-]Steven Byrnes3y40

Thanks! But this isn’t currently too important to me. I researched & wrote this in a couple hours and that’s about the most I want to spend on it. So I’ll pass.

Hey, if anyone reading this cares more than I do to get to the bottom of this issue, please DM ChristianKI per the above comment, and tell me if you wind up learning anything interesting. :)

[-]Andy_McKenzie3y20

Thanks for this good post. A meta-level observation is that people are grasping at straws like this is evidence that our knowledge of the causes of schizophrenia is quite limited.

[-]Yitz3y21

This is an excellent post, and has convinced me to stop repeating that “statistic” without some significant caveats

[-]Bamse2y10

I have not read all the

[-]Bamse2y-10

I think cortical blindness has a consequense for developing schizophrenia. Research has shown that that there are early disturbance in the visual cortex and retina at early schizophrenia. If you do not have a visual cortex you can not get the disease as it starts there.

Staffan

Claims & Evidence

So then Leivada & Boeckx 2014 narrowed the hypothesis to: congenital cortical blindness is protective against schizophrenia, congenital peripheral blindness is not. (Terminology in footnote.^[1]) That seems like an odd and unjustified direction for people to have pivoted. The problem is: congenital cortical blindness is even more uncommon than congenital blindness overall—that paper says 18% of congenitally blind people are cortically blind, while the data in Morgan et al. 2018 implies that it’s ≈10%. If the authors only had a handful of total reports of congenitally-blind schizophrenics in the first place, and if only 1 out of 5-10 congenitally-blind people are more specifically cortically-blind, then we really shouldn’t be shocked that none of that handful of documented cases is cortically-blind. That seems like the kind of thing that could easily happen by coincidence.

OK, leaving aside the false “it has never happened” claim, and the unjustified “cortical blindness” topic, we can retreat to a weaker form of the hypothesis: comorbidity of schizophrenia and congenital-blindness has happened, but it happens much less often than if they were statistically independent. Do we have evidence on this claim?

The best kind of evidence would be a population-wide systematic study. Unfortunately, the null hypothesis (schizophrenia and congenital-blindness are independent) has them co-occurring at a rate of 0.72% × 0.03% = 2 in a million. So you need a really big population to get good evidence on this hypothesis. Morgan et al. 2018 grabbed data from Western Australia, but it turned out (in my assessment) that they didn’t have enough data to come to a conclusion one way or the other.^[2] Jefsen et al. 2020 grabbed Danish data and found the same (non) conclusion.

Well, so much for that plan! Alternatively, we can look at how many published case reports of congenitally-blind schizophrenics there are, and compare it to how many we would have normally expected. If there are many fewer than expected, that’s evidence that blindness protects against schizophrenia. See Silverstein et al. 2013 for a nicely-laid-out argument that this is the case.

Again, if schizophrenia and congenital-blindness were independent, something like 0.72% × 0.03% = 2 in a million people would have it, i.e. 650 people in the USA. But the number of case reports is (they say, although see above) 0. I’ll admit it: 650 is a lot more than zero. But not everyone with schizophrenia gets properly diagnosed. (I think the 0.72% is total incidence, not how many people are diagnosed, but I could be misreading, here’s the source.) And there’s a much much bigger problem: not every psychiatrist, when treating a blind schizophrenic person, will publish a case report about them! Y’know, people are busy! Jeez!!

So, what’s the probability that any given blind schizophrenic patient will wind up immortalized in the literature as a case report? “Much less than 1 in 620” does not immediately strike me as implausible. I’m not an expert, of course. I don't know how easy and routine it is for psychiatrists & psychologists to publish case reports. Maybe every neighborhood therapist is just publishing case reports left and right, whenever they have a spare moment. Beats me! I tried to look up how many psychology case reports get published per year, but couldn't easily figure it out, please comment if you know.

What about unpublished anecdotes? Well, as soon as I looked, I did in fact quickly find a couple people on the internet offering anecdotal reports of congenitally-blind schizophrenics—quora example, reddit example. Granted: people sometimes lie on the internet! But for what it’s worth, the quora person claims to be a “DeafBlind Tech Trainer, ASL Interpreter, CVI Consultant” and claims to have met at least three blind schizophrenics “including congenitally blind people”, and this person has a bunch of other quora posts that seem pretty good and serious and consistent with their claimed identity—it’s not just a trolling account, or if it is, it’s an unusually high-effort trolling account!

Conclusion

My tentative conclusion is that, if I have a good way to think about schizophrenia, and everything hangs together well, except that it can’t explain why congenital / early blindness is protective against schizophrenia, I think I should stick with the theory and declare that congenital / early blindness is not actually protective against schizophrenia after all!

To be perfectly clear, if I had two equally-good-in-every-way theories of schizophrenia, and Theory A predicted that congenital blindness was somewhat protective against schizophrenia, and Theory B didn’t, I would consider that a point in favor of Theory A. It would just be a pretty small point in favor of Theory A.

^{^}

If I understand correctly, “cortical blindness” means that you’re blind because there’s something wrong with your occipital lobe, and “peripheral blindness” means you’re blind because there’s something wrong with your eye, retina, etc.

^{^}

This is my assessment. The authors definitely didn’t put it that way. As far as I can tell, the authors did no statistical analysis whatsoever; the only nod to statistics is a throwaway comment that their study is “possibly underpowered” (!). OK fine, let me get out my calculator. If I’m doing the math right, they expected to find 2.5 people in the early-peripherally-blind schizophrenia group and actually found 1, which is a p-value of 0.29; and they expected 0.264 people in the early-cortically-blind schizophrenia group and actually found 0, which is p-value of 0.77. If that’s right, then consider me underwhelmed! But please someone double-check my math!

LESSWRONG
LW

LESSWRONG
LW

83

I’m mildly skeptical that blindness prevents schizophrenia

83

83

Claims & Evidence

Conclusion