Background: Bipolar disorder (BD) is characterized by persistent disturbances in emotional regulation and cognitive function, even during euthymia, yet its neural mechanisms remain unclear. Given evidence of structural and dynamic functional connectivity (dFC) abnormalities in BD, investigating dynamic structure-function connectivity coupling (SC-FC coupling) may provide insights into the pathophysiology of the disorder. Methods: Diffusion tensor imaging and resting-state fMRI data were collected from 61 euthymic BD-I patients and 54 healthy controls (HC). Structural connectivity was derived using probabilistic tractography, and dFC was estimated with a sliding-window approach. SC-FC coupling was quantified as the extent to which structural features predicted dFC using multivariate linear regression. Abnormal SC-FC coupling in BD was assessed at both the network and regional levels, relative to HC. Over time, SC-FC coupling was categorized into high- and low-coupling states, and dynamic state analyses of dwell time, occurrence rate, and transition probabilities were performed in the affected brain regions. Results: Network-level analyses indicated that BD exhibits lower SC-FC coupling in subcortical regions compared with HC. Regional analyses revealed significantly reduced dynamic SC-FC coupling in two thalamic subregions (medial posterior and rostral temporal) and in the right primary somatosensory cortex in BD compared to HC. Dynamic analyses further demonstrated thalamic instability: BD showed shorter dwell times in high-coupling states, increased occurrence of low-coupling states, and more frequent high-to-low coupling transitions than HC. These abnormalities suggest that BD is characterized by disrupted structure-function integration and unstable thalamocortical dynamics. Conclusion: Our findings reveal regional abnormalities in SC-FC coupling dynamics in BD, indicating persistent dysfunction of structure-function integration during euthymia. Dynamic coupling measures may provide preliminary insights into potential biomarkers for disease monitoring.
Ye et al. (Wed,) studied this question.