Identification and Analysis of Key Players and Key Connections in Brain Networks

Abstract

The Key Player Problem (KPP), introduced by Stephen P. Borgatti, aims to identify a set of k nodes (KP set) whose removal maximally disrupts communication within a network. These disruptions can impair the integrity, strength, or efficiency of connections. While the KPP has been applied to various complex networks, its use in brain networks remains largely unexplored. In brain networks, such disrup tions are common in aging, neurological disorders, tumors, and brain injuries. However, in brain networks, edge removal representing the loss or weakening of connections between regions—is more prevalent. These edges may represent anatomical pathways or functional links. To address this research gap, we introduce the Key Connection Problem (KCP), defined as identifying a set of k edges (KC set) whose removal most severely disrupts network communication. We developed two algorithms Exhaustive k-Node Removal for Global Efficiency Minimization (EnGEM) and Exhaustive k-Edge Removal for Global Efficiency Minimiza tion (EnGEM-E) to detect both Key Players and Key Connections in a network. For the application of these algorithms, structural brain networks were constructed from neuroimaging data of 100 cognitively normal older adults (NC) and 100 Alzheimer’s disease (AD) subjects, each comprising 80 nodes and edges denoting anatomical pathways. Key Player analysis revealed that in normal aging, the most disruption-causing nodes were predominantly left-lateralized subcortical and memory related regions, whereas in Alzheimer’s disease, they shifted toward the right hemisphere and frontal regions, reflecting disease-related changes in network vulnerability. Key Connection analysis revealed that normal aging networks rely on posterior and limbic connections, while Alzheimer’s disease networks show disrupted frontal–subcortical connections, reflecting disease-specific structural connectivity alterations. This is the first study to formally define and investigate the KPP and KCP in brain networks, providing a new framework for analyzing structural connectivity dis ruptions in clinical neuroscience.