Section: New Results
Population shrinkage of covariance (PoSCE) for better individual brain functionalconnectivity estimation
Estimating covariances from functional Magnetic Resonance Imaging at rest (rfMRI) can quantify interactions between brain regions. Also known as brain functional connectivity, it reflects intersubject variations in behavior and cognition, and characterizes neuropathologies. Yet, with noisy and short timeseries, as in rfMRI, covariance estimation is challenging and calls for penalization, as with shrinkage approaches. We introduce population shrinkage of covariance estimator (PoSCE) : a covariance estimator that integrates prior knowledge of covariance distribution over a large population, leading to a nonisotropic shrinkage. The shrinkage is tailored to the Riemannian geometry of symmetric positive definite matrices. It is coupled with a probabilistic modeling of the individual and population covariance distributions. Experiments on two large rfMRI datasets (HCP n=815, CamCAN n=626) show that PoSCE has a better biasvariance tradeoff than existing covariance estimates: this estimator relates better functionalconnectivity measures to cognition while capturing well intrasubject functional connectivity.

More information can be found in [20].