Topology of Surface Displacement Shape Feature in Subcortical Structures

Amanmeet Garg, Donghuan Lu, Karteek Popuri, Mirza Faisal Beg

Abstract: The shape of anatomical structures in the brain has been adversely influenced by neurodegenerative disorders. However, the shape feature covariation between regions (e.g. subfields) of the structure and its change with disease remains unclear. In this paper, we present a first work to study the topology of the surface displacement shape feature via its persistence homology timeline features and model the polyadic interactions between the shape across the subfields of subcortical structures. Specifically, we study the caudate and pallidum structures for Shape Topology change with Parkinson’s disease. The shape topology features show statistically significant group level difference and good prediction performance in a repeated hold out stratified training experiment. These features show promise in their potential application to other neurological conditions and in clinical settings with further testing on larger data cohorts.