A Generative Model for Activations in Functional MRI

Detection of brain activity and selectivity using functional magnetic resonance imaging (fMRI) provides unique insight into the underlying functional properties of the brain. We propose a generative model that jointly explains neural activation and temporal activity in an fMRI experiment. We derive an algorithm for inferring activation patterns and estimating the temporal response from fMRI data, and present results on synthetic and actual fMRI data, showing that the model performs well in both settings, and provides insight into patterns of selectivity. Thesis Supervisor: Polina Golland Title: Associate Professor