Neuro-inspired Information Processing

3Neurons and Synapses

The cortex and its cortical columns are composed of two major cell categories: the nerve cells, also known as ‘neurons’, and a set of cells known as the glial cells. The glial cells are ten times greater in number than the neurons, but they are, in principle¹, incapable of emitting electrical signals. More particularly, they contribute to keeping the neurons alive, providing nutrients and oxygen, and they participate in the proper functioning of synapses and therefore in their essential property, which is to transmit electrical signals from one neuron to another. Little is known about their functions at present, but there is no doubt that they are essential to the overall functionality of the nervous system. However, since they are not directly involved in the generation and propagation of action potentials, we will not consider them in the rest of this chapter, which will focus on neurons and synapses.

3.1. Background

Before studying the mathematical and electrical models of neurons and synapses, let us offer a quick overview of some of their main characteristics and properties that will need to be reproduced when designing and producing artificial neural networks.

3.1.1. Neuron

We owe the discovery of neurons to Santiago Ramón y Cajal in 1887. Using a silver staining technique proposed by Camillo Golgi, known as ‘Golgi staining’, Ramón y Cajal showed that the brain was not a continuous medium, but that it was non-homogeneous and composed of a large number ...