Computational Neuroscience: Capturing the Essence

Abstract

The nervous system faces a most challenging task – to receive information from the outside world, process it, to change adaptively, and to generate an output – the appropriate behavior of the organism in a complex world. The research agenda of computational neuroscience (CN) is to use theoretical tools in order to understand how the different elements composing the nervous system: membrane ion channels, synapses, neurons, networks, and the systems they form, implement this demanding challenge successfully. CN deals with theoretical questions at both the cellular and subcellular levels, as well as at the networks, system, and behavioral levels. It focuses both on extracting basic biophysical principles (e.g., the rules governing the input-output relationship in single neurons) as well as on high-level rules governing the computational functions of a whole system, e.g., “How is a spot of light moving in the visual field encoded in the retina?” Or “how do networks of interconnected neurons represent and retain memories?” Ultimately, CN aims to understand, via mathematical theory, how do high-level phenomena such as cognition, emotions, creativity, and imagination, as well as brain disorders such as autism and schizophrenia, emerge from elementary brain-mechanisms. Here we highlight a few theoretical approaches used in CN and provide the respective fundamental insights that were gained. We start with biophysical models of single neurons and end with examples for models at the network level.