RAND Statistics Seminar Series

A Stimulus-Locked VAR Model for fMRI Event-Related Connectivity Analyses

Presented by Wesley K. Thompson, University of Pittsburgh
Thursday February 21, 2008
1:30 P.M. ET
Conference Room 6202
RAND Corporation, Pittsburgh, PA
Please contact Denise Miller if you would like to attend this seminar.


Functional magnetic resonance imaging (fMRI) enables neuroscientists to measure spatially-localized brain activation over time in response to experimental stimuli. In recent years neuroscientists have become increasingly interested in using fMRI to explore dynamic relationships (effective connectivity) among brain regions. Questions involving effective connectivity center on how functionally-specialized brain regions interact with each other in processing mental stimuli, and how these interactions differ among subgroups. One class of methods which have been proposed for effective connectivity analyses involves vector autoregressive (VAR) modeling of fMRI time series (Harrison, et al., 2003). To date, these VAR models have not been specifically tailored to the analysis of nonstationary fMRI data. This is a crucial issue because the slow event-related fMRI experimental designs which are best adapted for effective connectivity analyses result in regional activations and effective connectivity relationships which are time and stimulus dependent. This talk proposes a flexible semiparametric state space VAR model which allows for data-driven estimation of brain activation curves and connectivity coefficients. The connectivity coefficents are allowed to vary smoothly as a function of time from stimulus onset. Connectivity coefficients are also allowed to differ across individual subjects and subgroups via a random effects formulation. This model is applied to data obtained from a psychiatric neuroscience experiment comparing differences in connectivity of pre-selected brain regions between clinically depressed and healthy never-depressed controls.

Speaker Bio

Dr. Wesley Thompson is an Assistant Professor of Statistics with a joint appointment in Psychiatry. Dr. Thompson's research interests include functional data analysis, dynamic modeling of multivariate longitudinal data, and applications to psychiatric research. He has a mentored career award from the NIMH centered on developing and applying functional and latent variable methodologies to research on depression and cardiovascular health.

Attending a Seminar

Other Locations/Times:
Washington, D.C. Conf. Rm. 4132: 1:30pm ET
Santa Monica Conf. Rm. 1226: 10:30pm PT

RAND visitors are welcome to attend and must RSVP at least one day prior to the seminar. To ensure your attendance please contact Denise Miller at dmiller@rand.org with your name, company (or university) affiliation, and national citizenship (for security purposes).

For parking and directions to RAND's Santa Monica office, please see: http://www.rand.org/about/locations/santa-monica.html.

For parking and directions to RAND's Pittsburgh office, please see: http://www.rand.org/about/locations/pittsburgh.html.

For further information and to be added to the mailing list contact Denise Miller at dmiller@rand.org.