Computational Neuroscience and Systems Biology

In 1952, Alan Hodgkin and Andrew Huxley published a paper that was the result of several years of experimentation on the axon of the giant squid. They had been measuring action potentials, a task made easier in the giant squid due to the large diameter of its axons (up to 1mm, compared to 1 micrometer, or millionth of a meter, in humans). Using a device (known as a voltage clamp) that allowed them to manipulate the voltage of the axon membrane and measure the resultant current that flowed through its ion channels, they developed a mathematical model that could be used to calculate current flow across excitable membranes. They won the Nobel Prize in 1963 for their work, and amazingly their equations are still used today in their original form.

This mathematical modeling of neuronal function might be considered the first historical step in the creation of a field that is known today as computational neuroscience. Computational neuroscience involves the translation of brain function into quantifiable models. This usually necessitates drawing from a number of different fields, such as neuroscience, cognitive psychology, electrophysiology, mathematics, and computer programming.

A recent article in PloS Computational Biology summarizes the history of computational neuroscience and examines the interaction of the field with another: systems biology. Systems biology is an approach to studying biology that emphasizes looking at a biological system as a whole. This is in contrast to a reductionist methodology, which involves breaking something down into its constituent parts in order to understand how it functions.

Biology has had to rely on reductionism for much of its history, simply because there has not been enough information to understand whole systems. Now, however, sub-fields like genomics and proteomics have led to drastic gains in the extensiveness of our knowledge of biological processes, allowing complex computational modeling of biological systems to occur for the first time.

As pointed out in the PloS article, however, these two fields that use computational methods to explore neuroscience and biology, respectively, are distinctly separate from one another, and have little interaction or overlap. Why is this?

One reason is that the information available to systems biology is much more comprehensive. Data like an entire genome is accessible to use in computational modeling. Neuroscience, on the other hand, usually has to take a more theoretical approach. For example, computational neuroscientists do a lot of work with neural network models. These models, however, are usually general examples and don’t attempt to mimic specific networks in the brain. At this point, accurate modeling of distinct networks is a little too ambitious of an endeavor. The disparity in the information available to the two fields has led to differences in methods and tools (e.g. the software used for modeling), which make integration of the areas even more difficult.

It seems, however, that this chasm between computational neuroscience and systems biology will eventually be abolished. At this point it may be unavoidable, as knowledge of neuroscience lags behind that of other biological areas for various reasons that range from the complexity of the brain to the history of our philosophical approach to studying it. But the understanding of biological processes like gene expression and protein synthesis that makes systems biology capable of large-scale modeling attempts will eventually lead to an improved elucidation of how the brain works. This will inevitably allow for the integration of computational neuroscience and systems biology. After all, the brain is a pretty important part of the overall system.


De Schutter, E., Friston, K.J. (2008). Why Are Computational Neuroscience and Systems Biology So Separate? PLoS Computational Biology, 4 (5), e1000078. DOI:10.1371/journal.pcbi.1000078