Despite considerable progress in understanding the biology of Parkinson's disease (PD), reliable biomarkers are still lacking. The combination of non-invasive neuroimaging data reflecting both functional and structural characteristics of the brain, clinical information, and other biologic measures provides an unprecedented opportunity for cross-cutting investigations that stand to gain a deeper understanding of PD. We present a Bayesian statistical modeling framework that incorporates imaging data from different modalities and yields classifications of subjects as either PD patients or healthy controls (HCs). We apply our model to data from magnetic resonance imaging (MRI), diffusion tensor imaging (DTI), resting-state functional MRI, and numerous clinical variables. We also consider a penalized regression approach to distinguish PD patients from HC subjects. In both cases, we demonstrate the ability to isolate neural characteristics that reflect accurate signatures of PD and that, upon further investigation, may serve as useful early stage PD biomarkers.
Copyright. All Rights Reserved.