I. Introduction
Multi-Voxel pattern analysis (MVPA) is an emerging approach for studying the relationship between cognition and brain activity measured by functional magnetic resonance imaging (fMRI). fMRI measures blood oxygenation level-dependent (BOLD) signal that arises from the interaction between blood flow (and blood oxygenation) and changes in neural activity [2]. In task-based studies, changes in neural activity are assumed to be experimentally induced, with fMRI time series data reflecting the response to stimuli at certain locations (i.e., voxels) across the brain. The hemodynamic response coupled to neural activity is slow (on the order of seconds) and systematic, allowing the observed BOLD signal change associated with a given stimulus (or stimulus block) to be adequately modeled by a canonical response function and summarized by a single value.