Simulation of neuronal death and network recovery in a computational model of distributed cortical activity.

OBJECTIVES: The authors utilize a model of activity-dependent neuronal plasticity to study the interplay between synaptogenesis, neuronal death, and neurogenesis on the resulting pattern of neuronal connectivity. DESIGN: A mathematical model of neuronal network activity was employed, with plasticity instantiated by an activity-dependent rewiring rule. In particular, the authors modeled a neural system as a collection of "nodes" (neural subsystems) connected by "links" (anatomical connectivity). Neuronal damage was simulated by deletion of nodes in this evolving network through either random or targeted attack. Neurogenesis was likewise simulated by insertion of new nodes with random connections. MEASUREMENTS: Local and global structural network properties were characterized using the metrics of local and global "efficiency," and network "reachability." RESULTS: Activity-dependent plasticity yields a network that is robust to random node deletion, with preservation of a "small-world" architecture, characterized by high local and global efficiency. In contrast, targeted deletion of central nodes leads to a drop in reachability and global efficiency, with a consequent loss of small-world properties. Simulated neurogenesis is able to compensate for this targeted cell loss even when rates of new cell formation are considerably slower than that of simulated cell death. CONCLUSIONS: The rapid growth of computational neuroscience enables to study the interplay between neuronal plasticity and cell death in computational models of brain network activity. Although the current simulations lack much of the rich physiology of real neuronal systems, they nevertheless allow us to make tentative hypotheses of the effects of neuronal lesions on the resulting neuroanatomical connectivity networks.