Neural network model for path-finding problems with the self-recovery property

The large-scale synchronization of neural oscillations is crucial in the functional integration of brain modules, but the combination of modules changes depending on the task. A mathematical description of this flexibility is a key to elucidating the mechanism of such spontaneous neural activity. We present a model that finds the loop structure of a network whose nodes are connected by unidirectional links. Using this model, we propose a path-finding system that spontaneously finds a path connecting two specified nodes. The solution path is represented by phase-synchronized oscillatory solutions. The model has the self-recovery property: that is, it is a system with the ability to find a new path when one of the connections in the existing path is suddenly removed. We show that the model construction procedure is applicable to a wide class of nonlinear systems arising in chemical reactions and neural networks.