Bangjie Wang

bangjiewang.bsky.social

25 August 2026 11:49

Excited to share our new preprint on how the cerebellum specializes for language even in the absence of contralateral neocortical inputs, with @aniladmello.bsky.social and our collaborators @evfedorenko.bsky.social, @gretatuckute.bsky.social, @hopekean.bsky.social! www.biorxiv.org/content/10.6... 1/

Blake Richards

tyrellturing.bsky.social

30 July 2026 11:47

Yeah, from talking to @rezashadmehr.bsky.social, I gather that's roughly right, but with one caveat: The cerebellum does most of its learning in the synapses onto the P-cells, which are more like the *middle* layer of a three layer network. So, there's still a need to send errors backwards.

Blake Richards

tyrellturing.bsky.social

29 July 2026 16:36

Super cool paper here, showing that the cerebellum has a credit assignment mechanism solving the weight transport problem and giving every P-cell the right error signal! But, could the neocortex be the same? Could it possibly be hard-wired like this? I suspect not... Mystery to solve! 🧠📈 🧪

joshcashaback.bsky.social

29 July 2026 11:42

Looking forward to reading!

Andrew D Wilson

adw.bsky.social

29 July 2026 11:42

I’ve been pondering this for a while. What constitutes an error, outside of a lab defined operant conditioning set up? How does anything in the brain know something is wrong?