TRACKING REGIONAL CONNECTIVITY DYNAMICS IN PATIENTS WITH MILD COGNITIVE IMPAIRMENT AND ALZHEIMER’S DISEASE

Yu Sun,Zhaojun Zhu,Haijing Niu,Ying Han

DOI: https://doi.org/10.1016/j.jalz.2016.07.025

2016-01-01

Abstract:AD is a neurodegenerative disease which has been described as disconnection syndrome as its disruptions of the synchronous exception and synchronization of spontaneous neural activity between brain networks. Previous studies assessed functional connectivity (FC) by the full time series of signals over the entire scan, while brain FC are dynamically connected [1]. It has been proposed that quantifying time-varying FC may provide insight into fundamental properties of brain networks. Here, we used resting-state functional near-infrared spectroscopy (fNIRS) [2] to examine FC dynamics in patients with MCI and AD. Resting-state fNIRS data of 24 AD, 27 MCI and 31 healthy controls (HCs) were collected. After data processing, a sliding-window correlation analysis [3] was performed on fNIRS time series to obtain dynamic network organization at rest. Functional connectivity fluctuation strength (FCFS) maps were calculated as a metric to quantify fluctuation strength of FC in brain networks and four spatially different connectivity groups (homotopic connectivity, long and short intrahemispheric connectivity and heterotopic connectivity) between AD, MCI and HCs. Regression analysis with age, gender and education level removed was used to examine between-group differences and the results were corrected for multiple comparisons. Finally, correlation analysis was used to determine the relationships between the results and cognitive measures. Compared to HCs, the MCI and AD showed much larger FCFS values in cortical FC (Fig. 1 and Table 1). There were significant different FCFS between AD, MCI and HCs in default mode network, frontoparietal network, ventral attention network and visual network with AD group showing the largest values in these networks (Fig. 2). Within the homotopic connectivity, long and short intrahemispheric and heterotopic connectivity, there were significant FCFS differences between groups (Fig. 3(A, B)). Finally, we observed significant negative correlations between the FCFS values and MMSE in default mode network and long intrahemispheric connectivity group, and between the FCFS values and MoCA in the long intrahemispheric connectivity group (P < 0.05, corrected) (Fig. 4).