Inference of structure and dynamics in cognitive cortical code

Summary

Date: 
March 9, 2015 - 11:00am
Location: 
NW 243
About the Speaker
Name: 
KiJung Yoon (University of Texas at Austin)

Grid cells, which display striking, periodic spatial patterns of activity in open fields, appear to respond substantially differently on 1D tracks. The multiple response fields of a cell are not periodically arranged, they exhibit a wide range of peak amplitudes, and their mean spacing is several times larger than in 2D environments. Despite the low-dimensional covariates of grid cell responses including an animal’s location and velocity that can be easily controlled or measured, the neural representation or fine-scale (< a few minutes) dynamics has not extensively been studied especially in 1D due to the enigmatic 1D response. We ask whether the 1D response is consistent with observations in 2D, or whether the system is in a different dynamical regime. In this talk, I will present novel ways of analyzing grid cell data and inferring underlying structure and dynamics of the cells to show that 1D responses appear consistent with slicing through an equilateral triangular grid, and that cells in the same putative network are well-described by parallel 1D slices.