Neural Inflexibility and Temporal Fragmentation of Sensory-Attention Networks in Migraine with Aura: A Hidden Markov Model Study

Migraine with aura (MWA) is thought to involve cortical spreading depression and abnormal large-scale network organization, yet the temporal dynamics of brain states during the interictal period remain insufficiently understood. This study applied a Hidden Markov Model (HMM) to resting-state functional MRI data from 70 patients with interictal MWA and 70 age- and sex-matched healthy controls to characterize recurrent dynamic brain states and their clinical relevance. Ten latent functional states were identified and grouped into four archetypal configurations: sensory-attention coupling, default mode dominance, executive control dominance, and attention-control linkage. Compared with healthy controls, patients with MWA showed significantly reduced mean dwell time in a sensory-attention coupling state characterized by somatomotor and ventral attention network engagement, together with altered state-specific connectivity involving the default mode, somatomotor, and ventral attention networks. Within the MWA group, longer dwell time in this aberrant sensory-attention state was positively associated with Migraine Disability Assessment scores, suggesting that persistence in a dysfunctional sensory-attention configuration may contribute to disease burden. Patients also exhibited increased fractional occupancy in attention-control states involving dorsal attention and frontoparietal networks. Transition probability analysis further revealed a restricted dynamic profile in MWA, with enhanced transitions among executive and attentional configurations, whereas healthy controls displayed a more distributed and flexible transition repertoire. These findings indicate that interictal MWA is characterized by temporal instability of sensory-attention integration and a shift toward high-demand attentional control states. HMM-derived dynamic metrics may provide quantitative neuroimaging markers for characterizing network inflexibility, hypervigilance, and disability-related brain dysfunction in MWA.