Computational Neuroscience: a Brief Overview

Computational and mathematical modeling is an increasingly useful approach for investigating the functionality of the nervous system. Though such modeling has been used for decades, recent advances in computational power and numerical techniques have greatly expanded its scope, with a corresponding increase in research activity. This paper presents a brief – and necessarily incomplete – review of methods and applications in computational neuroscience. Computational neuroscience refers to the use of mathematical and computational models in the study of neural systems. It is part of the larger – increasingly active – discipline of computational biology, which applies computational modeling to all aspects of biological organisms. Quantitative modeling has been a key component of research in neuroscience for many decades. Indeed, one of the most celebrated achievements in the field – the Hodgkin-Huxley model for the generation of action potentials – was a triumph of the quantitative approach. Also, much of what is understood about the functionality of the visual, auditory and olfactory systems, as well as the neural basis of learning and memory, has been informed by mathematical and computational modeling. Nevertheless, it is fair to say that, until recently, computational modeling represented only a small part of the total research effort in neuroscience, which has traditionally been dominated by experimental studies. This has begun to change for several reasons which are discussed in the next section. The recent move towards computational modeling has opened up new directions of research, and allowed investigation of issues beyond those that are accessible to direct experimental study. More importantly, it has brought new ideas from fields such as statistical physics, information theory, nonlinear systems theory and engineering into neuroscience, providing a richer conceptual framework for answering the most difficult fundamental questions in the field. This overview discusses the motivation for the use of computational modeling in neuroscience, briefly describes some of the approaches, and looks at a few areas where these approaches have been fruitful. Several excellent texts providing details of methods and applications in computational neuroscience are now available.