Abstract
Objective
To characterize gray and white matter structural alterations in migraine without aura (MwoA) and migraine with aura (MwA) relative to healthy controls (HC), test direct differences between the migraine groups, and examine spatial associations between voxel-based morphometry findings and molecular and cellular reference templates.
Background
Structural and diffusion magnetic resonance imaging (MRI) studies have reported gray and white matter abnormalities in migraine, but findings vary across cohorts, imaging methods, and clinical subgroups. Whether migraine with and without aura show distinct structural patterns remains uncertain, and the molecular and cellular context of these alterations is incompletely understood.
Methods
This cross-sectional group-comparison analysis was embedded in an ongoing prospective single-center study at The First Affiliated Hospital of Dalian Medical University in Dalian, China, and included participants recruited from August 2025 to March 2026. Thirty HC, 40 participants with MwoA, and 15 participants with MwA underwent T1-weighted and diffusion MRI. Diffusion tensor imaging, neurite orientation dispersion and density imaging, and diffusional kurtosis imaging metrics were analyzed using tract-based spatial statistics; gray matter microstructure and volume were assessed using gray matter–based spatial statistics and voxel-based morphometry (VBM). JuSpace Toolbox correlated unthresholded VBM statistical parametric t (spmT) maps with neurotransmitter- and cell-related templates. Models adjusted for age, sex, and education; VBM additionally adjusted for total intracranial volume; and direct MwoA-versus-MwA sensitivity models included Patient Health Questionnaire-9.
Results
Relative to HC, MwoA showed higher mean and radial diffusivity and lower neurite density index, whereas MwA showed lower free-water fraction (all threshold-free cluster enhancement–family-wise error [FWE], p < 0.05). No diffusional kurtosis imaging or corrected gray matter–based spatial statistics result was significant. VBM showed lower gray matter volume in MwoA in left cerebellar Crus I (77 voxels; peak-level whole-brain family-wise error (FWE)-corrected p value (pFWE) = 0.022) and right inferior frontal gyrus, pars triangularis (6 voxels; pFWE = 0.025), and in MwA in right inferior frontal gyrus, pars triangularis (85 voxels; pFWE = 0.019). Direct MwoA-versus-MwA comparisons were nonsignificant in primary and Patient Health Questionnaire-9–adjusted models. JuSpace identified seven and 11 false discovery rate (FDR)-significant neurotransmitter-related associations for MwoA-versus-HC and MwA-versus-HC, respectively; no cell-related template survived FDR correction for MwoA-versus-HC, whereas nine survived for MwA-versus-HC (q < 0.05). No region of interest–clinical association survived Bonferroni correction.
Conclusion
Multimodal MRI identified gray and white matter alterations in both migraine groups relative to HC, whereas direct MwoA-versus-MwA comparisons were nonsignificant. JuSpace provided complementary spatial context for the VBM findings. These results broaden the structural characterization of migraine but do not establish imaging differences between the two migraine groups.
