TITLE: Modeling of BrainMap data

AUTHORS: Finn Årup Nielsen and Lars Kai Hansen

Department of Mathematical Modelling, Building 321
Technical University of Denmark, DK-2800 Lyngby, Denmark
emails: fn,lkhansen@imm.dtu.dk
www: http://eivind.imm.dtu.dk


We apply machine learning techniques in the form of Gaussian mixture models to functional brain activation data. The dataset was extracted through the WWW interface to the BrainMap(tm) (Research Imaging Center, University of Texas Health Science Center at San Antonio) neuroimaging database. Modeling of the joint probability structure of activation foci and other database entries (e.g. behavioral domain, modality) enables us to summarize the accumulated body of activation coordinates in the form of a 3D density and allows us to explore issues like the effect of the experimental modality on the resulting brainmap.

Submitted for NIPS'99, Denver, Colorado, 1999.