4/ The brain cared about these variables too. Lesions and Ca2+ imaging showed secondary motor cortex (M2) represented recent actions, checking behavior, reward, and elapsed time, a much richer representation of history than we usually include in simple trial-by-trial models.
love a green lynx spider
1/ We work hard to factor "noise" out of behavior. But the brain doesn't.
Strip away noise and you may miss what the brain evolved to do
I wanted to share this pre-Bluesky paper where we found premotor (M2) corticostriatal circuits encode a broad history of behavior, beyond just action + reward 👇
(also love the Bursera species "Mexican Frankincense" it's prowling along - the leaves smell amazing)
8/ Ultimately, wanting to study a richer, more natural behavior is what pulled me toward another self-generated behavior built on cortico-basal ganglia circuits: birdsong. But even here, there's room to incorporate richer social and contextual variables...
6/ Optogenetic suppression of M2→DMS during actions, but not before or after, prevented use of recent experience.
That timing showed the circuit mattered when history had to be converted into action
3/ This gave us rich experiential variables to model. Prior experience wasn't just action + reward: checking behavior and elapsed time scaled how much recent history mattered
And mice who used that experience performed better
2/ While constrained tasks are useful, animals have evolved to use diverse sources of info to guide their behavior. We tried to move a small step in this direction by creating an un-cued, self-paced, self-guided lever hold down task
5/ M2's projection into dorsal medial striatum (M2→DMS) carried a narrower signal: recent lever pressing + checking behavior.