Title: Dimension reduction methods for multilevel neural firing rate data
Authors: Angel Garcia de la Garza - Columbia University (United States) [presenting]
Jeff Goldsmith - Columbia University (United States)
Abstract: Recent advances have allowed high-resolution observations of firing rates for a collection of individual neurons; these observations can provide insights into patterns of brain activation during the execution of tasks. Our data come from an experiment in which mice performed a reaching motion following an auditory cue, and contain measurements on firing rates from neuron in the motor cortex before and after the cue. We focus on appropriate dimension reduction techniques for this setting, in which sharp increases in firing rates after the cue are expected; we also allow for correlation across neurons in each of the trials, and within a neuron across trials. Initial results suggest differing patterns of activation, perhaps representing the involvement of different motor cortex functions at different times in the reaching motion.