From Functional Brain Models to Imaging Data

Neural circuits will be modeled with population density methods. In general, neural circuits are complex and unwieldy, due to the large number of parameters that need to be chosen only a few of which can be determined experimentally. Since the complex of individual neurons can be highly complex already, understanding neural circuits that are comprised by tens of thousands of neurons is difficult. Population density methods reduce the problem in two ways: the individual neural model is reduced to two dimensions, and neural circuits are reduced to populations. Modelling neural circuits as two-dimensional dynamical systems subject to noise simplifies the description considerably, but allows for enough flexibility to capture the rich dynamics demonstrated by more biophysically motivated models.

We will make large-scale models of neural circuits, in cortex and in spinal cord. In cortex, we will contribute to simulations in an existing simulator, The Virtual Brain, which is currently used to find explanations for epileptic seizures. We will add simulations that model the neural state, and not just the observed electrophysiological signal. In spinal cord, we will model the neural circuits that form a Central Pattern Generator and the feedback loops to and from muscles, with the aim of explaining observed EMG signals (Chakarbarty lab, FBS, Leeds). Ultimately, we hope to be able to model disturbed function of the spinal cord, e.g. cerebral palsy, and help to explore strategies to alleviate these conditions, informed by our models.

The project will entail part theoretical development of population level modeling, part development of novel numerical strategies to create simulation techniques, part simulation. Our approach falls broadly under the 'Healthcare Technologies theme, and specifically addresses mathematical biology and non-linear sciences.

The research will also be carried out in alignment with the theory group SP4 of the Human Brain Project, as part of the sub-task bridging scales. The supervisor is PI in this project. One PhD student will be funded by that project and will also work on the development of the theory of simulating neural populations. The implications for this project are substantial networking and dissemination opportunities, and a strengthening of the local group of which Hugh is a part.