Computational neurobiology

Bridging the gap from molecular structure to human behaviour. We will use novel approaches for collecting data at different spatial and temporal scales to develop multi-scale models that improve understanding of brain function in health and diseases.

Understanding how the brain works is one of the great challenges in the life sciences and is crucial for developing better approaches to treat brain disorders such as epilepsy, autism and schizophrenia.

In addition to these urgent medical needs, understanding fundamentals of brain function will also assist in solving many current engineering problems, such as building fault tolerant circuits; developing low energy computation; and building smart machines that can deal with novel situations.

While the brain has been the subject of intense and fruitful scientific investigation, there are still enormous gaps in our understanding. For example, it is still not clear on which level of brain organisation some of our most fundamental behaviours appear. We know memories can be stored, yet are they stored in the molecular properties of single neurons, small groups of local neurons, or in networks that span larger areas? Computational neurobiology can help address this fundamental challenge by formalising the investigation of brain function and creating mathematical models based on experimental observations. A good model does not only describe the data used to create it, but can also predict how the system will behave when faced with a novel set of circumstances. Computational neurobiology can span the biological and physical sciences providing the language to describe and understand the behaviour of one of the most complex devices known to man, the human brain.

Current Project: Epilepsy

Recent advances in the genetics of epilepsy have led to the discovery of hundreds of mutations that cause seizure syndromes. Almost without exception these mutations are in ion channels — molecules that reside in the cell membrane and conduct current that is the basis of electrical signalling in the nervous system. In the Computational Neurobiology Laboratory, we use a variety of techniques to understand how these molecules are different from those in humans without epilepsy.

One such technique is to artificially grow these ion channels in special cell lines and measure the differences in their electrical properties compared to ion channels from people without epilepsy. These differences are small and subtle, and it is often unclear how they cause the brain to become prone to seizure. One way to make the casual connection between molecular deficit and changes in brain dynamics is through the use of computer models. Data collected from cell line studies can be converted to mathematical models of the ion channel in question. These models can then be incorporated into models of single neurons and used to predict differences in sensitivity of these neurons to interactions from other neurons. These neuronal models can be incorporated into network models with thousands of neurons to help us understand what triggers seizures. Understanding why some brains are more prone seizures than others will enable us to develop better treatments but also tell us why brains become epileptic in the first place.

Laboratory leader

To find out more about the research undertaken at this laboratory, please contact: