Kota, John, Kerry discussed on Brain Inspired

Brain Inspired
|

Automatic TRANSCRIPT

Many neurons collected at the same time to get some better snapshot of the population has been productive especially in the past five or ten years not just a representation questions that have to Cognition as john pointed out evening in motor cortex has been tremendously fruitful. So you have on the one hand. A relatively stagnant research program that uses these older techniques and has a commitment to a sharing tony in style. Explanation where you're detailing individual neurons and the connections between them and what each individual neuron is. Doing i've heard it described as sort of like this Anecdotal neuroscience where each neuron has a little tail to tell about what it does to one where you have. Many neurons are being recorded That have produced a novel insight into a range of of cognitive as well as non cognitive neural functions. So i think and that's productive so unlike this adherence to this more old-fashioned view there has been a productive research program and and that to me says not just that this is the new thing in town but this is sort of what the doctor ordered that we really did need these kinds of techniques and methods in order to make progress. And that's very exciting. I mean this is. This is actually really good news. And i'll add as by way of kota here that it is not just confirming these models that we have from other disciplines. Either it's really providing novel insight into how the brain is able to rise to the intelligent behavior that's truly representation cognitive. Can you give a country example of a of a novel insight. Yeah so a lot of the recent research and publications that are coming out the have to do with this kind of new hop fielding into where you're looking rural populations in you're looking at how the system traverses through that state space have provided insight into more complex representations in the brain a law and one of the main explanatory concepts as notion of a metaphor. So it's not like when you look at these population activity and you have thousands of neurons. Let's say they're being recorded that the system occupies any point in this huge space of possibilities instead. The system tends to occupy points. That are on what's called a manifold Sort of a restricted subset of very restricted subset of the possible positions in that space. And it's those subsets are manifold that kerry seemed to carry these kinds of representations contents And there's there's nothing about any theories in psychology or connor science and saying you have to use manifold or anything like this but the brain does use manifold. So i'd like to think of like to conceptualize this kind of insight as a novel contribution. I'll add in addition that there are ways of reading out where the system lies in the restricted sub spaces on these manifold that is also providing novel insight into cognition. So i i love this one paper. It's a beautiful paper From the jazayeri lab at mit. It's from sewn. At all in the journal neuron from two thousand eighteen where they were investigating how the brain is able to keep track of and recreate Intervals of time. So the animals have to wait through an interval time. And then when they're signaled to recreate it they kind of have to forego moving they have to inhibit their movements until they think. Unequal amount of time has elapsed at which time they respond and in addition to these kinds of manifold i.

Coming up next