Edge‐centric functional network reveals new spatiotemporal biomarkers of early mild cognitive impairment

Weiping Wang,Ruiying Du,Zhen Wang,Xiong Luo,Haiyan Zhao,Ping Luan,Jipeng Ouyang,Song Liu

DOI: https://doi.org/10.1002/brx2.35

2023-01-01

Abstract:Abstract Most neuroimaging studies of the pathogenesis of early mild cognitive impairment (EMCI) rely on a node‐centric network model, which only calculates correlations between brain regions. Considering the interaction of low‐order correlations between pairs of brain regions, we use an edge‐centric network model to study high‐order functional network correlations. Here, we compute edge time series (eTS) to obtain overlapping communities and study the relationship between subnetworks and communities in space. Then, based on the overlapping communities, we calculate the normalized entropy to measure the diversity of each node. Next, we compute the high‐amplitude co‐fluctuation of the eTS to explore the pattern of brain activity with temporal precision. Our results show that the normal control and EMCI patients differ in brain regions, subnetworks, and the whole brain. In particular, entropy values show a gradual decrease, and brain network co‐fluctuation increases with disease progression. Our study is the first to investigate the pathogenesis of EMCI from the perspective of spatiotemporal flexibility and cognitive diversity based on high‐order edge connectivity, further characterizing brain dynamics and providing new insights into the search for biomarkers of EMCI.