Neuroimaging-based subtyping of migraine identifies clinically distinct phenotypes

Area of Science:

• Neuroimaging
• Neurology
• Brain Imaging

Background:

• Migraine heterogeneity is not fully understood.
• Integrating structural and functional brain imaging may reveal neurobiological differences.
• Previous studies have not comprehensively explored multimodal imaging for migraine subtyping.

Purpose of the Study:

• To identify distinct migraine subgroups using a data-driven, multimodal neuroimaging approach.
• To characterize the clinical and imaging profiles of these subgroups.
• To compare multimodal clustering with unimodal (functional or structural only) approaches.

Main Methods:

• 111 individuals with migraine underwent resting-state functional connectivity (FC) and structural MRI scans.
• Principal component analysis reduced data dimensionality, followed by hierarchical agglomerative clustering.
• Two multimodal subgroups (M1 and M2) were identified and compared with controls and unimodal clusters.

Main Results:

• Multimodal clustering identified two distinct migraine subgroups (M1 and M2).
• M2 exhibited older age, longer disease duration, greater disability, increased widespread FC, and reduced cortical volumes compared to M1 and controls.
• M1 showed preserved cortical structure and stronger control-network connectivity, with no significant deviations from controls.

Conclusions:

• Data-driven multimodal neuroimaging successfully identified two distinct migraine subgroups.
• These subgroups differ in clinical characteristics and brain structure/function, highlighting migraine heterogeneity.
• This approach provides a framework for imaging-informed migraine characterization and further research.