Modeling Approaches for Multiple Types of Brain Signals

• Activity Number: 143
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Raquel Prado*+
• Companies: University of California, Santa Cruz
• Keywords: Bayesian time series ; brain connectivity ; state-space models

Abstract

Clinical and non-clinical neuroscience research often involves joint recording of different types of brain signals, such as functional magnetic resonance imaging (fMRI) and electroencephalogram (EEG) signals. Each data modality provides information about brain activity in a specific spatio-temporal resolution. For instance, fMRI data have high spatial resolution and low temporal resolution, while EEGs have high temporal resolution and low spatial resolution. Therefore, analyses that consider both types of signals can lead to better inference on the spatio-temporal characteristics of brain activity. We consider hierarchical Bayesian state-space modeling approaches for estimating brain connectivity using different types of signals recorded in multi-trial and multi-subject studies. We illustrate the performance of the models and methods in the analyses of synthetic and real brain data.