A1242
Title: Reliability for covariance matrix-valued data: Application to resting stat fMRI
Authors: Yu Huang - Columbia University (United States)
Philip Reiss - University of Haifa (Israel)
Seonjoo Lee - Columbia University/New York State Psychiatric Institute (United States)
Todd Ogden - Columbia University (United States) [presenting]
Abstract: Resting-state functional MRI (rs-fMRI) data, which is used to measure intrinsic brain connectivity, can be naturally represented as covariance matrices that reside on a Riemannian manifold. Theoretically grounded reliability assessments are described for such data that leverage geometric frameworks, employing a Riemannian metric to quantify stability and reproducibility. Using this framework, design considerations in rs-fMRI studies are investigated, including a selection of regions of interest, scan duration, and inter-scan intervals. Through application to the midnight scanning club dataset, it is demonstrated how these geometric approaches provide actionable insights for optimizing rs-fMRI protocols while accounting for the inherent manifold structure of connectivity data.
