Blindness, Schizophrenia, and Autism

A short psychiatric-trivia post (⚠ carries Scott's later 'all of this might be false' update). The striking facts: no congenitally blind person gets schizophrenia, yet congenitally blind children are ~50x more likely to be autistic (~50% meet criteria). Scott connects this to the diametrical autism-vs-schizophrenia model (opposite corpus-callosum size, nicotine use, rubber-hand-illusion susceptibility, self-other boundaries) and offers his preferred speculative synthesis: some neural hyperparameter is sensitive to total sensory input, and losing vision shifts it toward the autism pole — which would neatly explain both the never-schizophrenia and the excess-autism, plus the subclinical autistic features in blind non-autistic kids. Points at Corlett/Frith/Fletcher predictive-coding work as the next step.

Quality 62 · Strong. Strong/solid edge: memorable, thought-provoking, and offers a genuine unifying speculative model over two real findings — but short, admittedly speculative, and flagged as possibly-debunked.

Claude’s paradigm shift 46 · Moderate. Notable: the sensory-input-shifts-a-hyperparameter frame is an original (if tentative) synthesis of the blindness/schizophrenia/autism facts.

Real-world impact 2 · Minor. Within-discourse: an interesting hook for computational-psychiatry readers; no material reach (and the finding may not hold).