MR Functional Connectivity as an Effective Prognostic Biomarker for Degree of PD Medication Reduction after DBS Parkinson’s Disease

Motivation: Parkinson's disease symptom response to deep brain stimulation (DBS) is highly variable. There is a strong need for more accurate and reliable prognostic criteria to guide treatment. Goal(s): Extract pre-operative motor network areas whose resting-state fMRI functional connectivity can strongly and robustly predict DBS outcomes, improving on existing clinical prognostic criteria. Approach: Feature extraction algorithms were implemented, particularly LASSO selector, to find functional connectivity pairs that can best predict patients' responses to DBS. Results: Cortico-striato-cerebellar connections could explain around 20% additional variance in DBS outcomes compared to traditional clinical predictors including pre-operative motor scores. Impact: We show that pre-operative measures of functional connectivity in motor regions relevant to Parkinson's disease mechanisms can explain differences in patient response to deep brain stimulation (DBS). These prognostic criteria can aid in filtering DBS candidates and improve patient-centric outcomes.