Joint estimation of multiple clinical variables of neurological diseases from imaging patterns

This paper presents a method to estimate multiple clinical variables associated with neurological pathologies from brain images, aiming to quantitatively evaluate continuous transition of neurological pathologies from the normal to diseased state. Built upon morphological measures derived from structural MR brain images, a Bayesian regression method is developed to jointly model multiple clinical variables for capturing their inherent correlations and suppressing noise. Coupled with a feature selection technique, the regression method is used to build a joint estimator of multiple clinical variables associated with Alzheimer's disease from structural MR brain images of elderly individuals. The cross-validation results demonstrate that the proposed method has superior performance over existing techniques.