Functionally informed cortex based alignment: an integrated approach for whole-cortex macro-anatomical and ROI-based functional alignment

Abstract

Due to anatomical variability across subjects many brain mapping experiments have analysis focused on a few particular regions of interest so as to circumvent the problem of sub-optimal statistics resulting from the lack of anatomical correspondence across subjects. Since the topographic distribution of experimental effects across the cortex is also often of interest, two separate analyses are often conducted, one on the regions of interest alone, as well as a separate 'whole brain' analysis with sub-optimal spatial correspondence across brains. In this paper we present a new group alignment procedure which incorporates, from each subject, both macro-anatomical (curvature) information and functional information from standard localizer experiments. After specifying appropriate parameters to weight anatomical and functional alignment forces, we were able to create a group cortical reconstruction which was well aligned in terms of both anatomical and functional areas. We observed an increase in the overlap of functional areas as well as an improvement in group statistics following this integrated alignment procedure. We propose that, using this alignment scheme, two separate analyses may not be necessary as both analyses can be integrated into a single procedure. After an integrated structural and functional alignment one is able to carry out a whole brain analysis with improved statistical sensitivity due to the reduction in spatial variation in the location of functional regions of interest which fCBA accomplishes. Furthermore, regions in the vicinity of localised and aligned regions-of-interest will also benefit from the integrated alignment.

Keywords: Group alignment; Group statistics; ROI analysis; Spatial normalisation; fMRI.