Investigating Heritability Across Resting State Brain Networks Via Heat Kernel Smoothing on Persistence Diagrams

Abstract

The brain’s heritable topological differences in resting state functional connectivity (rsfc) measured via resting state fMRI (rsfMRI) provide important insight into brain function and dysfunction. Current techniques investigating heritability are limited by arbitrary rsfc threshold selection and reduction of otherwise detailed brain topological properties into summary measures. Topological Data Analysis (TDA) is a novel tool for addressing these limitations by analyzing how the topological properties of data vary without arbitrary threshold and summary metric construction. TDA applies a filtration to the data and constructs a persistence diagram (PD). Therefore, the purpose of this study was to compute PDs to determine TDAbased heritability of static brain network topological features. To this end, we calculated a robust heritability index map across smoothed PDs derived from twin rsfMRI data.