Linking cortical morphology and neurophysiological dynamics in Parkinson's disease

Parkinson's disease (PD) involves progressive neurodegeneration and distinctive structural and functional alterations in cortico-basal ganglia circuits. This study proposes bridging the gap between structural and functional biomarkers to uncover fundamental mechanisms underlying PD pathophysiology and support more comprehensive diagnostic and therapeutic approaches. This study analyzed intraoperative electrocorticography (ECoG) and globus pallidus internus (GPi) local field potential (LFP) recordings alongside preoperative T1-weighted structural MRI from 50 patients with PD undergoing deep brain stimulation surgery. We extracted 36 morphometric features (thickness, volume, surface area) from nine sensorimotor Brodmann areas using FreeSurfer and 92 neurophysiological features (power, burst dynamics, coherence) across multiple frequency bands. Pairwise analyses revealed only fragmented, though significant, correlations. In contrast, the sparse partial least squares (SPLS) analysis identified a robust and significant latent dimension (test set rho = 0.818, p = 0.001) linking the two modalities. This primary latent variable was driven by a strong negative association with cortical thickness in sensorimotor areas (e.g., BA3b, BA1, BA6) and a complex combination of neurophysiological features, most notably altered burst dynamics in the alpha, gamma, and low-beta bands. This structure-function relationship was independent of age, disease duration, and UPDRS-III (unified Parkinson's disease rating scale) scores (rho = 0.701 after partialling out confounds). These findings highlight the value of multimodal approaches for uncovering structure-function interactions in PD, and the potential of integrated biomarkers for improving diagnosis and treatment strategies.