Multimodal 7T MRI reveals subcortical structural, functional, iron accumulation, and glymphatic abnormalities in chronic migraine

Background Exploring the pathogenesis of chronic migraine (CM) remains a formidable challenge. While 7T MRI enables precise detection of microstructural alterations in subcortical nuclei and subregions in CM, comprehensive multimodal imaging studies are few. We therefore conducted a 7T multimodal imaging study to characterize subcortical nuclei and limbic subregions within the hypothalamus, thalamus, hippocampus, and amygdala alterations associated with migraine chronification and to determine the impact of medication-overuse headache (MOH) on CM brain changes. Methods We performed multimodal 7T MRI in 62 healthy controls (HCs), 62 episodic migraine (EM) patients, and 62 CM patients (31 with MOH). MRI measures included volumes, functional indicators (fractional amplitude of low-frequency fluctuation (fALFF) and regional homogeneity (ReHo)), quantitative susceptibility mapping (QSM), and diffusion tensor imaging analysis along the perivascular space, choroid plexus volume, and basal ganglia perivascular space (PVS) metrics for glymphatic function. Group differences and associations with clinical measures were assessed. Results After false discovery rate (FDR) correction, fALFF in the accumbens area (L) and mean individual PVS volume differed significantly among the three groups ( q < 0.05). Both EM and CM showed lower left accumbens fALFF and greater PVS volume than HCs, whereas no significant difference was detected between EM and CM. No group effects survived FDR correction for regional volume, ReHo, QSM, or other PVS metrics. No significant differences were found between CM patients with and without MOH. Left accumbens fALFF was negatively correlated with the headache duration, whereas PVS volume was positively correlated with headache duration. Conclusions Migraine was associated with reduced spontaneous activity in the left accumbens area and increased basal ganglia PVS volume. The absence of detectable differences between EM and CM suggests that these abnormalities may represent shared migraine-related features rather than changes that increase linearly with chronification. The findings implicate altered pain-reward processing and perivascular morphology, while indicating that the transition to CM may depend on changes that static regional measures do not capture.