Multilevel Structure-Function Coupling Reveals Network Signatures of Remission in Paroxysmal Kinesigenic Dyskinesia.
Authors
Affiliations (7)
Affiliations (7)
- Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China.
- Institute of Medical Robotics, Shanghai Jiao Tong University, Shanghai, China.
- Department of Neurology/Genetics and Rare Diseases, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China.
- Department of Neurology, Shanghai General Hospital Affiliated to Shanghai Jiaotong University School of Medicine, Shanghai, China.
- Shanghai Professional Technical Service Platform for Genetic and Rare Neurological Diseases, Shanghai, China.
- Neurological Disorder Center, Haikou Orthopedic and Diabetes Hospital of Shanghai Sixth People's Hospital, Haikou, China.
- National Engineering Research Center of Advanced Magnetic Resonance Technologies for Diagnosis and Therapy, Shanghai Jiao Tong University, Shanghai, China.
Abstract
Paroxysmal kinesigenic dyskinesia (PKD) causes brief, movement-triggered dystonic or choreic attacks that can severely affect daily life. Though most patients eventually remit, timing and rate of improvement differ substantially across individuals, and predictive biomarkers remain lacking. Evidence indicates that PKD reflects distributed network dysfunction. To investigate how structural wiring supports or constrains these dysfunctions, as evaluated by structure-function (structural connectivity-functional connectivity [SC-FC]) coupling. Diffusion kurtosis imaging and resting-state functional magnetic resonance imaging (fMRI) were obtained in 95 patients with PKD and 44 healthy controls (HC). SC-FC coupling was quantified at nodal, intranetwork, and internetwork levels. Patients were classified as remission or nonremission. Machine-learning classifiers based on multilevel SC-FC features were trained to distinguish PKD from HCs and to predict remission status. Associations between SC-FC features and disease duration were assessed using Spearman's correlation. Patients with PKD exhibited widespread SC-FC coupling abnormalities consistent with distributed rather than focal network dysfunction. Remission and nonremission subgroups exhibited distinct SC-FC signatures, centering on cerebellar, somatomotor, and default-mode systems. Machine-learning classifiers discriminated PKD patients from HCs (area under the receiver operating characteristic curve [AUC] = 0.91) and remission from nonremission patients (AUC = 0.95). Cerebellar-default-mode network coupling was the top discriminative feature and correlated positively with disease duration (Spearman's r = 0.26, P = 0.031, FDR corrected). Multilevel SC-FC coupling analyses reveal systems-level abnormalities in PKD and suggest that partial normalization of coupling patterns accompanies remission. Cerebellar-default-mode network coupling may serve as a sensitive imaging biomarker of remission status and disease progression, highlighting coupling-based metrics as candidates for predicting PKD trajectory. © 2026 International Parkinson and Movement Disorder Society.