Treatment of vascular cognitive impairment (VCI) in adult moyamoya disease (MMD) is still unclear because of its unveiled neural synchronization. This study introduced a dynamic measurement of connectivity number entropy (CNE) to characterize both spatial and temporal dimensions of network interactions. Fifty-one patients with MMD were recruited (27 with VCI and 24 with intact cognition), as well as 26 normal controls (NCs). Static network properties were first examined to confirm its aberrance in MMD with VCI. Then, the dynamic measurement of CNE was used to detect the deteriorated flexibility of MMD with VCI at global, regional, and network levels. Finally, dynamic reconfiguration of flexible and specialized regions was traced across the three groups. Graph theory analysis indicated that MMD exhibited "small-world" network topology but presented with a deviating pattern from NC as the disease progressed in all topologic metrics of integration, segregation, and small-worldness. Subsequent dynamic analysis showed significant CNE differences among the three groups at both global (p < 0.001) and network levels (default mode network, p = 0.004; executive control network, p = 0.001). Specifically, brain regions related to key aspects of information processing exhibited significant CNE changes across the three groups. Furthermore, CNE values of both flexible and specialized regions changed with impaired cognition. This study not only sheds light on both the static and dynamic organizational principles behind network changes in adult MMD for the first time, but also provides a new methodologic viewpoint to acquire more knowledge of its pathophysiology and treatment direction.
Keywords: Functional flexibility; Moyamoya disease; Network topology; Resting-state fMRI; Vascular cognitive impairment.