Shape Analysis Using Spherical Harmonic Description with Point Distribution Models (SPHARM-PDM): A Powerful Tool For Investigating Morphological Brain Changes

Background: Pathology within subcortical structures is implicated in various psychiatric and neurological disorders. Associated with a loss of function is the loss of form, which can be used as biomarker of disease onset or progression. While volumetric changes can provide a global measure of subcortical structure alteration, this does not take into account regional changes. Applying shape analysis to subcortical structures allows more precise delineation of localised structural changes. Objectives: We aimed to examine the localised structural changes in subcortical structures across a variety of disease states and to correlate changes in form and function. The process, advantages and challenges of shape analysis are discussed. Methods: The caudate, putamen, nucleus accumbens and thalamus were manually segmented from magnetic resonance imaging in patients with Parkinsons disease, Huntingtons disease, behavioural-variant frontotemporal dementia and bipolar disorder. Shape analysis, using the spherical harmonic description with point distribution models (SPHARM-PDM) package, is then applied to these segmentations to determine localised deformations in the surface of the structures. Examples are provided across these various diseases that highlight the associations between function and form. Findings: Shape analysis provides a complement to volumetric analysis, providing insights into regional changes in subcortical structures. Atrophy in subregions frequently correlates with known subcortical circuits. Nevertheless, challenges with determining intersubject structure correspondence still remain. Conclusions: Shape analysis is a powerful tool in investigating in vivo morphological brain changes. Moreover, the Australian US Scandinavian Spanish Imaging Exchange (AUSSIE) network provides a shared platform to examine these changes across a wide variety of pathophysiologies and to refine shape analysis techniques in the future.