Encoding and decoding models of functional MRI responses in human visual and auditory cortex.

Abstract

Computational models of brain activity serve to not only uncover the organizational principles of the brain, but also quantify the neural processes underlying our perceptual experiences. This dissertation investigates the neural encoding of sensory information through computational models of blood oxygen level-dependent (BOLD) responses obtained with functional magnetic resonance imaging (fMRI). Based upon a technique known as population receptive field (pRF) modeling, voxel-based encoding and decoding models were used to predict and decode neural responses to various visual and auditory stimuli. The studies presented herein demonstrate the utility of this technique for examining the functional organization of the visual and auditory cortex, as well as provide methods for verifying model generalizability across different stimulus types.