State-specific disruptions of dynamic functional connectivity in young migraine without aura: a hidden Markov model approach.

BACKGROUND

Migraine is a common neurological disorder involving network-level dysfunction. Increasing evidence suggests that migraine involves network-level dysfunction and is associated with altered resting-state functional connectivity. Traditional static functional connectivity analyses are limited in capturing the temporal dynamics of large-scale brain networks. The Hidden Markov Model (HMM) provides an advanced analytical framework to identify discrete, recurrent brain states and characterize their temporal properties without the constraints of arbitrary windowing assumptions.

OBJECTIVE

To characterize dynamic functional connectivity alterations in young patients with migraine without aura (MWoA) using HMM and examine associations between dynamic state metrics and clinical disability.

METHODS

Resting-state fMRI data were obtained from 200 participants (100 young MWoA patients and 100 matched healthy controls). Using the Dosenbach 160 ROI template (cerebellum excluded; N = 142), HMM identified recurring brain states. Group differences in fractional occupancy (FO), mean dwell time (MDT), and state transition probabilities were assessed. State-specific functional connectivity patterns were compared, and correlations with clinical indices (MIDAS, VAS, HIT-6) were evaluated.

RESULTS

Eleven robust dynamic brain states were identified. Compared with controls, migraine patients demonstrated increased FO and MDT in States 7 (dorsal attention network-dominant) and 9 (multisensory integration), alongside reduced values in sensorimotor states (States 3, 4, 8, 11). State 9 exhibited significant abnormalities in DMN-SC and DMN-VAN connectivity (FDR-corrected q < 0.05). Transition analyses revealed enhanced self-transitions and increased incoming transitions to States 7 and 9, whereas transitions among sensorimotor states were diminished. MDT in State 9 was positively correlated with MIDAS scores (r = 0.38, p < 0.05), indicating its association with functional disability.

CONCLUSIONS

Young MWoA patients exhibit a dual-mode dysfunction in brain dynamics: heightened external vigilance (State 7) and impaired segregation of internal-external processing (State 9), accompanied by instability in baseline sensorimotor configurations. Prolonged dwelling in State 9 and its correlation with disability highlight this multisensory integration state as a potential biomarker of migraine-related functional impairment. These findings provide new insights into neurobiological mechanisms and support dynamic network-based therapeutic strategies.