Neurodegenerative diseases have been shown to target large-scale neural networks. Experimental data show that intrinsic connectivity in healthy subjects predicts region-by-region vulnerability to disease. Thus, specific regions play the role of critical network “epicenters,” whose normal connectivity profiles showed a prominent similarity to the disease-associated atrophy pattern. Standard graph connectivity analyses in healthy subjects have determined that regions with higher total connectional flow and shorter functional paths to the epicenters, are more prone to neurodegenerative diseases. These analysis techniques only measure symmetric (undirected) connections between regions, and cannot differentiate between direct and indirect links, nor infer causality.
Other studies have shown that a high interconnection between brain modules facilitates the development of network failures, as do the so-called bridge nodes, which are regions belonging to many modules. At the same time, most studies are based on analyzing the connectome of healthy older subjects upon which it is believed that neurodegeneration is superimposed. However, the ideal approach would be to follow subjects from health to neurodegenerative disease, detecting vulnerable connectivity sources and interactions within single subjects. This would require a shift in the current paradigm of analysis of the connectomics of brain disease from static graph networks, based on graph theoretical measures, to dynamic graph networks that employ novel theoretical control mechanisms and can be applied to elucidate the above phenomena. Most prominent concepts are leader-follower networks, distributed winner-take-all networks, detection of leader nodes via pinning observability/controllability and graph distances, and area aggregation and time-scale modeling.
This Research Topic solicits papers that respond to this challenge and overcome modeling obstacles encountered with current techniques, and develop and adapt emerging control strategies in graph theory for the much-needed understanding of neurodegenerative disease networks’ dynamic properties and the processes responsible for disease evolution. We seek new descriptors of complex networks able to quantify induced changes in topology or network organization, or to serve as theory-driven biomarkers to be used in disease prediction at the level of the individual subject.
Neurodegenerative diseases have been shown to target large-scale neural networks. Experimental data show that intrinsic connectivity in healthy subjects predicts region-by-region vulnerability to disease. Thus, specific regions play the role of critical network “epicenters,” whose normal connectivity profiles showed a prominent similarity to the disease-associated atrophy pattern. Standard graph connectivity analyses in healthy subjects have determined that regions with higher total connectional flow and shorter functional paths to the epicenters, are more prone to neurodegenerative diseases. These analysis techniques only measure symmetric (undirected) connections between regions, and cannot differentiate between direct and indirect links, nor infer causality.
Other studies have shown that a high interconnection between brain modules facilitates the development of network failures, as do the so-called bridge nodes, which are regions belonging to many modules. At the same time, most studies are based on analyzing the connectome of healthy older subjects upon which it is believed that neurodegeneration is superimposed. However, the ideal approach would be to follow subjects from health to neurodegenerative disease, detecting vulnerable connectivity sources and interactions within single subjects. This would require a shift in the current paradigm of analysis of the connectomics of brain disease from static graph networks, based on graph theoretical measures, to dynamic graph networks that employ novel theoretical control mechanisms and can be applied to elucidate the above phenomena. Most prominent concepts are leader-follower networks, distributed winner-take-all networks, detection of leader nodes via pinning observability/controllability and graph distances, and area aggregation and time-scale modeling.
This Research Topic solicits papers that respond to this challenge and overcome modeling obstacles encountered with current techniques, and develop and adapt emerging control strategies in graph theory for the much-needed understanding of neurodegenerative disease networks’ dynamic properties and the processes responsible for disease evolution. We seek new descriptors of complex networks able to quantify induced changes in topology or network organization, or to serve as theory-driven biomarkers to be used in disease prediction at the level of the individual subject.
