Diffeomorphic Alignment of Along-Tract Diffusion Profiles from Tractography

Along-tract diffusion measure profiles from white matter fiber tracts have been widely analyzed in population studies of neurodevelopment and disease. An implicit assumption in performing inter-subject comparisons is a one-to-one correspondence of the profiles across subjects. Further, the profiles may also exhibit noise and spatial variability due to misregistration, tractography algorithms, interpolation methods, or inter-subject differences in the population. We present an approach to minimize the variability in the shape of along-tract diffusion profiles by performing diffeomorphic alignment across the tracts as well as across the population. The method represents the tract profiles as configuration functions and defines an objective function to align configurations by minimizing over both global and local warping functions. Following alignment, we show decrease in variability using the measures of coefficient of variability and intraclass correlation coefficients. We also introduce a new measure of inter-tract correlation, obtained by correlation of diffusion profiles between a pair of tracts of interest. We demonstrate our methods in a population of healthy individuals as well as show reliability analysis in subjects scanned twice.