The key question in our study is, do monkeys point the joystick towards where the prey IS, or towards where they prey will soon be? That is, do they PROSPECT and then use that prospection to guide behavior?https://www.nature.com/articles/s41593-019-0561-6 …
-
-
Prikaži ovu nit
-
The monkey’s avatar's speed limit is faster than the prey’s speed limit, so both strategies are guaranteed too work, but prospection gives the reward a bit faster, and is therefore incentivized. It's a lot more intuitive once you watch the video:https://www.haydenlab.com/pursuit
Prikaži ovu nit -
Result is that monkeys DO prospect. We trained n=3 just to be sure. All three subjects headed towards the likely position of the prey in about 700-800 ms. Interestingly they all started out not prospecting, but then improved with a few weeks of practice, then stabilized.pic.twitter.com/MUelhsGrf9
Prikaži ovu nit -
OK, but do they actually prospect? That is, do they form an explicit neural representation of the prey's future position? Or just use some mindless heuristic strategy? (Gigerenzer aficionados know about the baseball players famous "mindless" predicting)https://www.annualreviews.org/doi/abs/10.1146/annurev-psych-120709-145346?casa_token=ugeWv35RNKMAAAAA:LYfOpHg-TBJb1PmPpPWtbsXuF38l58styv8VjohFEX8BRSseqXBxGOCoEoaqCWSNRhS4Q__x3Oc …
Prikaži ovu nit -
Prospection was the central philosophical question addressed by a series of
@templeton_fdn "Science of Prospection" Workshops that I attended and benefited greatly from.@neuromochi and I decided that in this case, answering the question required neuroscience.Prikaži ovu nit -
We asked, does the brain carry an explicit representation of the anticipated future position of the prey? Answer: Yes. More specifically, neurons in the dorsal anterior cingulate cortex (dACC) maintain an explicit continuously updated map of anticipated future prey position.
Prikaži ovu nit -
The map is a kind of place-cell-but-more-blobby representation, akin to non-grid medial entorhinal cells. Indeed, we used a GLM approach developed by Hardcastle,
@SuryaGanguli, and@lisa_giocomo for this purpose. https://www.sciencedirect.com/science/article/pii/S0896627317302374 … Here's a typically cell from our data :pic.twitter.com/k79CSElfpZ
Prikaži ovu nit -
We have much more to say about the mapping properties of dACC, including evasion as well as pursuit here.https://www.biorxiv.org/content/10.1101/796375v1 …
Prikaži ovu nit -
Moving forward, we are interested in going beyond simple choice tasks and towards dynamic on-line decisions. That means throwing out microeconomic theory and replacing it with control theory, and throwing out the concept of value and replacing it with control-theoretic variables.
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
Very cool work! Speculative thought: If the prey were of different sizes/values, you could -- with multiple targets -- do very high speed assessment of sensitivity to delay. Do monkeys always/sometimes opt for a smaller & closer target over a bigger further one?
-
This is a very good question. There’s a complication though which is that there is likely a control cost associated with changing trajectory. Just looking at the data - and doing the task ourselves - that cost looks quite high relative to the cost of time. We are currently
- Još 5 drugih odgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.