Multimodal Remote Digital Phenotyping for Detecting and Tracking Early Parkinsonian Change in LRRK2 Carriers

Identifying prodromal Parkinson's disease among LRRK2 carriers is critical yet challenging. We present a remote, multimodal video framework analyzing 829 participants, including 158 LRRK2 carriers (36 with manifest PD, 122 non-manifest), to address two key challenges: detecting high-risk carriers prior to clinical diagnosis and monitoring early disease-related change. Our AI model distinguished non-manifest carriers from controls with 92.9% accuracy (AUROC 0.92, AUPRC 0.82). Furthermore, our continuous PD Weigh-In score captured clinical decline in two carriers who subsequently developed PD and correlated strongly with expert ratings (Pearson r = 0.77, Spearman ρ = 0.79) across the held-out LRRK2 test cohort.