Kenneth Harris

University College London
Session: computational infrastructure for neuroscience: automation / pipelines

 

High-Dimensional Geometry of the Cortical Population Code as Revealed by 10,000-Cell Recordings
We used 2-photon calcium imaging and improved analysis methods to record the responses of >10,000 neurons in the visual cortex of awake mice, to thousands of natural images. The recorded population code was high-dimensional, with the variance of its dimensions following a powerlaw. This powerlaw did not reflect the statistics of natural images, as it persisted even when presenting spatially whitened stimuli.A mathematical analysis showed that a powerlaw signifies the maximum decorrelation possible for rate vectors lying on a smooth manifold, and predicted a 1+2/d relationship between the powerlaw exponent and the dimensionality of the input space. We confirmed this prediction experimentally.