Lin Huang, Xu Huang, Yuanwen Yu, Chuanyong Yu, Yu Wang
CM is associated with more extensive and frequency-specific CEN-related neuromagnetic abnormalities than EM, particularly within prefrontal-frontoparietal circuits. These findings suggest altered executive-control-related cortical dynamics during migraine chronification, while combined spectral and connectivity features may help differentiate CM from EM.
BACKGROUND: Chronic migraine (CM) is more disabling than episodic migraine (EM), but the neurophysiological mechanisms underlying migraine chronification remain incompletely understood. We investigated central executive network (CEN)-related neuromagnetic abnormalities in CM and EM.
METHODS: The original cohort included 22 patients with CM, 20 with EM, and 25 healthy controls (HCs). An independent validation cohort included 30 patients with CM and 30 with EM. Source-space analyses focused on 10 predefined CEN-related regions from the Desikan-Killiany atlas. Relative power spectral density (PSD) across six frequency bands and functional connectivity using corrected amplitude envelope correlation and network-based statistics were assessed. An exploratory support vector machine (SVM) analysis evaluated the discriminative value of CEN-related spectral and connectivity features.
RESULTS: PSD abnormalities were mainly observed in the theta, beta, and gamma1 bands and were largely driven by CM. Compared with HCs, CM showed increased theta-band relative power in left frontal CEN-related regions and reduced beta- and gamma1-band relative power in right frontal regions. Compared with EM, CM showed increased theta-band relative power in left frontal regions. Functional connectivity analysis revealed frequency-specific subnetworks in the theta, beta, delta, and gamma1 bands. Compared with HCs, CM showed increased theta- and gamma1-band connectivity and reduced beta-band connectivity, whereas EM showed more limited reductions in beta- and gamma1-band connectivity. CM also showed stronger delta- and gamma1-band connectivity than EM, predominantly involving prefrontal regions. The SVM combining two PSD and eight connectivity features achieved an accuracy of 90.48% and an AUC of 0.9773 (95% CI: 0.93-1.00) in the original cohort, and an accuracy of 86.67% and an AUC of 0.8322 (95% CI: 0.76-0.95) in the independent validation cohort.
CONCLUSION: CM is associated with more extensive and frequency-specific CEN-related neuromagnetic abnormalities than EM, particularly within prefrontal-frontoparietal circuits. These findings suggest altered executive-control-related cortical dynamics during migraine chronification, while combined spectral and connectivity features may help differentiate CM from EM.