Altered morphological connectivity mediated white matter hyperintensity-related cognitive impairment

Haifeng Chen,Jingxian Xu,Weiping Lv,Zheqi Hu,Zhihong Ke,Ruomeng Qin,Ying Chen,Yun Xu

DOI: https://doi.org/10.1016/j.brainresbull.2023.110714

2023-10-01

Abstract:White matter hyperintensities (WMH) are widely observed in older adults and are closely associated with cognitive impairment. However, the underlying neuroimaging mechanisms of WMH-related cognitive dysfunction remain unknown. This study recruited 61 WMH individuals with mild cognitive impairment (WMH-MCI, n = 61), 48 WMH individuals with normal cognition (WMH-NC, n = 48) and 57 healthy control (HC, n = 57) in the final analyses. We constructed morphological networks by applying the Kullback-Leibler divergence to estimate interregional similarity in the distributions of regional gray matter volume. Based on morphological networks, graph theory was applied to explore topological properties, and their relationship to WMH-related cognitive impairment was assessed. There were no differences in small-worldness, global efficiency and local efficiency. The nodal local efficiency, degree centrality and betweenness centrality were altered mainly in the limbic network (LN) and default mode network (DMN). The rich-club analysis revealed that WMH-MCI subjects showed lower average strength of the feeder and local connections than HC (feeder connections: P = 0.034; local connections: P = 0.042). Altered morphological connectivity mediated the relationship between WMH and cognition, including language (total indirect effect: -0.010; 95 % CI: -0.024, -0.002) and executive (total indirect effect: -0.010; 95 % CI: -0.028, -0.002) function. The altered topological organization of morphological networks was mainly located in the DMN and LN and was associated with WMH-related cognitive impairment. The rich-club connection was relatively preserved, while the feeder and local connections declined. The results suggest that single-subject morphological networks may capture neurological dysfunction due to WMH and could be applied to the early imaging diagnostic protocol for WMH-related cognitive impairment.