←
Sanjay Manohar582 posts
S

Sanjay Manohar

@BrainInTheMind
◷ Joined February 2012▧ 27 Photos and videos
Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·Nov 6

* What makes code good? * Today, perfecting your programming style is more important than ever. I'm thrilled that my book on Good Coding Practice is out after 13 yrs of work! It's written for any programming language, with examples in Matlab, Python & R. global.oup.com/academic/prod…

Post attachment
Oxford Academic avatar
⟳ Sanjay Manohar retweeted
Oxford Academic@OUPAcademic·Jan 6

Our latest Oxford Biology Primer by Sanjay Manohar @braininthemind is a concise introduction to best practices for coding that everyone who uses code in scientific research should know about. Find out more: bit.ly/4p1B1vh

Post attachment
bioRxiv Neuroscience avatar
⟳ Sanjay Manohar retweeted
bioRxiv Neuroscience@biorxiv_neursci·21 Dec 2025

Why motor learning involves multiple systems: an algorithmic perspective biorxiv.org/content/10.64898… #biorxiv_neursci

Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·14 Nov 2025

It also simulates human ERP responses to grammar. Overall, our rapid plasticity rules capture the structure-content duality of thought, unlike other neural networks. 6/6

Post attachment
Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·14 Nov 2025

The model runs in continuous time, and can encode words presented at variable speed. The synaptic rules predict lexical and syntactic priming seen in humans. It can learn syntax by very slow plasticity at the same synapses. Both order and morpheme-based grammars work. 5/6

Post attachment
Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·14 Nov 2025

Patients with damage to Broca’s area are often agrammatic: their speech is meaningful but lacks grammar. In contrast, patients with Wernicke’s area damage often have grammatical but nonsense speech. Specific model lesions can recreate these disorders. Unlike LLMs! 4/6

Post attachment
Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·14 Nov 2025

Like humans, the network recalls grammatical sentences better. How? Long-term synapses specify each word’s roles, and which role orders are OK. On this background, synapses change rapidly, creating a dynamic attractor for the sentence. 3/6

Post attachment
Sanjay Manohar avatar
Sanjay Manohar@BrainInTheMind·14 Nov 2025

We built a tiny untrained network (5 word and 5 role neurons) to encode and repeat a sentence. 100% explainable! Two rapid synaptic rules: 1) to associate words with their syntactic roles (subject, verb etc.), and 2) to retain the sequence of roles. 2/6

Post attachment
Gheorghe Iuga avatar
⟳ Sanjay Manohar retweeted
Gheorghe Iuga@gheorgheiuga·13 Sep 2025

Probably the coolest use of sin and cos functions you have ever seen!

ア@yuruyurau13 Sep 2025

a=(y,d=mag(k=(y<5?6+sin(y^1)*6:4+cos(y))*cos(i+t/4),e=y/3-13)+sin(e/4-t)/3)=>point((q=y*k/5*(2+sin(d*2+y-t*4)))+90*cos(c=d/3-t/2+i%2)+200,q*sin(c)+d*29-170) t=0,draw=$=>{t||createCanvas(w=400,w);background(9).stroke(w,96);for(t+=PI/90,i=1e4;i--;)a(i/790)}//#つぶやきProcessing

0:12