Gray–white matter structural signatures across migraine subtypes characterized by multimodal MRI and JuSpace correlation analysis

Background: Migraine is a common neurological disorder with a substantial disabling burden, yet its pathophysiological mechanisms remain incompletely understood. Previous MRI and diffusion MRI studies have reported gray and white matter abnormalities in migraine, but many were based mainly on conventional diffusion metrics, focused on a single subtype or tissue compartment, and rarely linked structural findings to molecular or cellular substrates. We integrated multimodal MRI with cross-modal spatial correlation analyses to evaluate gray and white matter structural alterations in migraine without aura (MwA) and migraine with aura (MA), and to explore their spatial associations with neurotransmitter and cell-surface marker maps. Methods: Thirty healthy controls (HC), 40 patients with MwA, and 15 patients with MA underwent 3D T1-weighted and diffusion MRI for DTI, NODDI, and DKI. White matter alterations were assessed using TBSS, gray matter microstructure using GBSS, and GMV using VBM. Mean values from significant TBSS and VBM clusters were extracted for ROI-based correlations. JuSpace analyses were performed on GM images showing between-group differences. Results: Compared with HC, TBSS revealed increased MD and RD and decreased neurite density index (NDI) in MwA, whereas in MA only free water fraction (FWF) survived correction (HC > MA); no DKI metric showed significant group differences after correction. Corrected GBSS analyses were negative, although uncorrected analyses suggested frontal-cingulate midline abnormalities. VBM demonstrated reduced GMV in MwA, mainly in the left cerebellar Crus I, and in MA, mainly in the right inferior frontal gyrus, pars triangularis; no significant GMV differences were detected between MwA and MA. JuSpace analyses based on VBM-derived difference maps showed significant spatial correlations with multiple neurotransmitter and cell-surface marker templates in both contrasts, with a broader set in MA vs HC. No ROI-based clinical correlation survived multiple-comparison correction. Conclusions: These findings support subtype-weighted structural phenotypes in migraine. MwA was characterized mainly by white matter microstructural abnormalities suggestive of altered neurite/axon-related tissue properties, whereas MA showed a more prominent free-water-related signal on the white matter skeleton. The spatial coupling between GMV alterations and neurotransmitter/cell templates further suggests that migraine-related structural changes may be embedded within a broader neurochemical and cellular context.