Statistical Modeling

Thomas D. Wickens
University of California at Berkeley

This talk introduces several issues in statistical analysis and modeling that are the basis
for techniques specifically used to analyze imaging data. I will begin by describing the
standard linear model of statistics, then look at its extensions to accommodate non-
Gaussian data (the Generalized Linear Model) and data in which the observations are
not independent. Finally, I will discuss the critical problems of multiple tests and the
selection of models.

Presentation (PDF File)

Back to Graduate Summer School: Mathematics in Brain Imaging