Investigating the Impact of Subject-Specific Region of Interest Definition in fMRI Analysis

Abstract

In the rapidly expanding field of neuroscience, functional magnetic resonance imaging (fMRI) is a widely used technique for neuroimaging. However, there are both anatomical and functional differences on an individual level that are often not accurately represented in fMRI analyses. The standard methods of analysis include defining the region of interest (ROI) either a priori or via average group response. The standard method of fMRI analysis requires a normalization which limits individual neural variability represented in the data. This thesis aims to investigate if using an individualized, subject-specific method of ROI definition increases the sensitivity of fMRI analysis. This question will be investigated by evaluating two experimental datasets using both a standard ROI definition and an individualized one, then comparing the statistical findings between the two methods for each dataset. Significant statistical differences were not found between the two methods in “think” versus “no think” trials nor in intrusion versus no intrusion trials.