Xiaohui Xie and memory
Xiaohui Xie came over from the Broad Institute to talk about memory.
He started by describing a typical computer memory circuit called a latch, which is a way of storing data using cross-coupled NAND gates. The outputs feed back into the inputs, maintaining the stored state. Then he talked about ants, and how they meander around the nest until they find food and then head directly back to the nest. Somehow the ants maintain their sense of orientation so they know where the nest is. Finally, he brought up eye position.
The way I understand it, there are two types of neurones involved in moving the eye: the saccade neurone and the position neurone. When the eye moves, the saccade neurones fire and then stop. The position neurones respond by changing how quickly they fire. Even though the saccade neurone has stopped firing, the position neurone remembers how quickly it is supposed to be firing. This is another type of memory. How do these position neurones keep firing if the eye stays in the same position for a long time? Neurones have a time constant of less than 10 milliseconds, but position neurones can fire for up to 10 seconds, how?
Xiaohui's point is that, like the computer memory circuit, the position neurons probably connect back to themselves in a positive-feedback loop. Positive feedback creates the persistence (memory) that locks the eye in place. This feedback system has to be finely tuned to prevent the eye from overshooting or drifting backwards. In fact, both of these symtoms have been seen by doctors in patients with eye problems.
He started by describing a typical computer memory circuit called a latch, which is a way of storing data using cross-coupled NAND gates. The outputs feed back into the inputs, maintaining the stored state. Then he talked about ants, and how they meander around the nest until they find food and then head directly back to the nest. Somehow the ants maintain their sense of orientation so they know where the nest is. Finally, he brought up eye position.
The way I understand it, there are two types of neurones involved in moving the eye: the saccade neurone and the position neurone. When the eye moves, the saccade neurones fire and then stop. The position neurones respond by changing how quickly they fire. Even though the saccade neurone has stopped firing, the position neurone remembers how quickly it is supposed to be firing. This is another type of memory. How do these position neurones keep firing if the eye stays in the same position for a long time? Neurones have a time constant of less than 10 milliseconds, but position neurones can fire for up to 10 seconds, how?
Xiaohui's point is that, like the computer memory circuit, the position neurons probably connect back to themselves in a positive-feedback loop. Positive feedback creates the persistence (memory) that locks the eye in place. This feedback system has to be finely tuned to prevent the eye from overshooting or drifting backwards. In fact, both of these symtoms have been seen by doctors in patients with eye problems.
0 Comments:
Post a Comment
<< Home