A physical view of computational neurodynamics

The nervous system is made of a large number of neurons. Time-varying balance between excitatory and inhibitory neurons is important to activate appropriate modes of electrical activity. A realistic biological neuron is complex, often presenting various electrophysiological activities and diffusive propagation of ions in the cell. Therefore, the physical effects of electromagnetic induction become very important and should be considered when estimating signal encoding and mode selection. Synaptic plasticity and anatomical structure have been developed to enhance the self-adaption of neurons. Thus, the electrical mode with the most effective links and weights can be selected to benefit information encoding and signal propagation between neurons in the network. As a result, the demand for metabolic energy can be greatly reduced. In this review, neuron model setting with biophysical effects, modulation of astrocytes, autapse formation and biological function, synaptic plasticity, memristive synapses, and field coupling between neurons and networks are reviewed briefly to provide guidance in the field of neurodynamics.