Assessment of the reproducibility of HARDI tractography based cortical connectivity measures suitable for clinical populations using a bootstrap approach

Abstract

Structural connectivity of the brain using MR diffusion tractography has gained significant interest. A connectivity matrix of cortical connectivity may provide unique insight into brain organisation. We aimed to develop a method to determine the number of seeds required to obtain stable and reproducible connectivity, and to assess reproducibility over time. We employ a bootstrap approach for estimation of these parameters. While connectivity measures of some regions are highly reproducible over time, other connections show poor reproducibility. This study highlights the relationship between seed number and reproducibility of connectivity.