A Multidimensional “Source‐Dynamics‐Function‐Structure” Magnetic Resonance Imaging Framework for Characterizing the Pathophysiological Imaging Profile of Moderate‐to‐Advanced Parkinson’s Disease Eligible for Deep Brain Stimulation

To characterize glymphatic‐ and cerebrospinal fluid (CSF)‐related imaging alterations in moderate‐to‐advanced, deep brain stimulation (DBS)‐eligible Parkinson’s disease (PD) using a multidimensional magnetic resonance imaging (MRI) framework.

We studied 60 moderate‐to‐advanced, DBS‐eligible patients with PD and 30 frequency‐matched healthy controls with 3.0 T MRI. The “Source‐Dynamics‐Function‐Structure” framework integrated choroid plexus volume fraction, C1–C2 phase‐contrast MRI, diffusion tensor image analysis along the perivascular space index, and perivascular space volumetry. Multivariable analyses were used to examine associations between imaging biomarkers and motor severity, anxiety, and sleep quality.

Patients with PD showed reduced diffusion tensor image analysis along the perivascular space index, a nominal increase in choroid plexus volume fraction, and a low‐net‐flow CSF pattern with nominal evidence of increased reflux and marked spatial heterogeneity, despite preserved gross perivascular morphology. Ventral CSF dynamics showed significant partial correlations with motor severity, whereas selected dorsal flow metrics remained significantly associated with anxiety severity in covariate‐adjusted regression analyses. The combined logistic model showed modest diagnostic performance (area under the curve = 0.686), with the highest area under the curve numerically but no statistically significant advantage over individual imaging metrics in paired DeLong comparisons.

This framework provides an exploratory neuroimaging reference for moderate‐to‐advanced, DBS‐eligible PD and may support future validation studies and DBS‐related investigations.