Quantifying spatial pattern similarity in multivariate analysis using functional anisotropy

Multivoxel pattern analysis (MVPA) has gained enormous popularity in the neuroimaging community over the past few years. At the group level, most MVPA studies adopt an “information based” approach in which the sign of the effect of individual subjects is discarded and a non-directional summary statistic is carried over to the second level. This is in contrast to a directional “activation based” approach which is typical in univariate group level analysis, in which both signal magnitude and sign are taken into account. The transition from examining effects in one voxel at a time vs. several voxels (univariate vs. multivariate) has thus tacitly entailed a transition from directional to non-directional signal definition at the group level. While a directional MVPA approach implies that individuals share multivariate spatial patterns of activity, in a non-directional approach each individual may have a distinct spatial pattern of activity. Here we show using an experimental dataset that indeed directional and non-directional MVPA approaches uncover distinct brain regions with some overlap. Moreover, we developed a descriptive measure to quantify the degree to which subjects share spatial patterns of activity. Our approach is based on adapting the Fractional Anisotropy (FA) measure, originally developed for examining diffusion MRI signals, in a novel way to quantify the degree to which subjects share a spatial pattern of activity. We term this measure “Functional Anisotropy” (FuA). Applying FuA to an auditory task, we found higher values in primary auditory regions compared to secondary and control regions. This highlights the potential of the FuA measure in second-level MVPA analysis to detect differences in the consistency of spatial patterns across subjects and their relationship to functional domains in the brain.