mathewsanders 5 years ago

> They mapped how the neurons in the ADN circuit fired as the animal’s head changed direction. Together, these data points formed a cloud in the shape of a simple and persistent ring.

Am I the only one who moved my head around and imagined my network activating to stay in sync?

It would be fascinating to know what inputs this network is using to determine current position (inner ear? sight? feedback from next muscles?) and doubly interesting to see what associated networks this ring is feeding into (I assume among other things it will be considering inner ear and head position to help maintain balance while walking)

> Further analysis revealed that the ring acts an attractor. If neurons stray off trajectory, they are drawn back to it, quickly correcting the system.

I’m at least a decade behind in neural net research, but have we gotten any closer to replicating this sort of system where a neural net can self correct over time? My mental model of an artificial neural net is fixed input/output equation that doesn’t have a function over time.