Quantitative multi-scale modelling of the human auditory cortex

Quantitative multi-scale modelling of the human auditory cortex
Modelling and Analytics for Medicine and Life sciences Doctoral Training Centre: PhD Scholarship

Supervisors: Professor Steve Coombes (School of Mathematical Sciences), Dr Chris Sumner (MRC Institute of Hearing Research), Patrick May (Leibniz Institute, Magdeburg), Dr Katrin Krumbholz (MRC Institute of Hearing Research)

Project description: Understanding how the underlying sensory processing by neural networks in the human brain gives rise to sensory perception is a difficult problem.

It is straightforward to measure human perception via behavioural experiments, but even accessing, let alone understanding, the underlying signals in the brain is difficult. Non-invasive measures such as EEG and fMRI allow monitoring of aggregate responses, but they are dramatically limited in their resolution, either spatially or temporally. Relating non-invasive responses back to underlying neural activity necessitates solving an inverse problem.

Moreover, our knowledge about how individual neurons behave can only be drawn from animal experiments. There is currently no principled way for inferring the behaviour of underlying neural circuits from non-invasive measurements.

This project will address the problem of sensory perception by developing a novel computational model of one part of the brain: the auditory cortex. This model will have the power to simulate individual spiking neurons, large populations of neurons, and far-field electrical signals (EEG, MEG) that are normally accessible in humans. This forward-model will allow the testing of hypotheses about the possible ways in which neural activity could give rise to the non-invasive observations, which in turn, are linked to results from behavioural experiments.

This will be achieved by bringing together two existing models: a large scale firing rate model of multiple auditory cortical fields, constrained by all the known anatomy (May et al. 2015); and a modelling framework which allows a rigorous abstraction from spiking models of neurons to neural field models (Byrne et al. in press). These field models can capture the dynamics of extended regions of the brain (Coombes 2010), be projected onto surfaces, and folded in the manner of the cortical surface. From these, far-field potentials (EEG, MEG) can be predicted. Thus for the first time we aim to provide a computational model that can predict, in a principled way, non-invasive measurements from the responses of single neurons. This will function as a platform for theoretically linking various measures of neural activity to sound perception.

References
Coombes S (2010). Large-scale neural dynamics: Simple and complex, NeuroImage, Vol 52, 731-739.
May PJ, Westoe J, Tiitinen H. (2015). Computational modelling suggests that temporal integration results from synaptic adaptation in auditory cortex. Eur J Neurosci. 41:615-30.
A Byrne, M J Brookes and S Coombes 2016 (in press). A mean field model for movement induced changes in the beta rhythm, Journal of Computational Neuroscience.

The MAML programme: The MAML doctoral training programme focuses on innovative modelling, simulation and data analysis to study real-world problems in medicine and biology. Maintaining a healthy society creates major challenges in areas including ageing, cancer, drug resistance, chronic disease and mental health. Addressing such challenges necessitates continuing development and implementation of a raft of new mathematical approaches and their integration with experimental and clinical science. Students will apply mathematical approaches (from areas such as dynamic modelling, informatics, network theory, scientific computation and uncertainty quantification) to research projects at the forefront of biomedical and life sciences identified through well-established collaborations with both academic and industrial partners.

MAML students will be provided with an excellent training environment within the Centre for Mathematical Medicine and Biology and collaborating departments. Students will undertake tailored training, complemented by broadening, soft-skills, wet-lab (where appropriate) and student-led activities. There will also be opportunities for training and exchanges with world-leading partners.


Applications: Please follow the instructions at the MAML website: http://www.nottingham.ac.uk/mathematics/maml Applicants for the MAML programme should have at least a 2:1 degree in mathematics, statistics or a similarly quantitative discipline (such as physics, engineering, or computer science).

Completed applications and references should be submitted by Wednesday 28 February 2018.


For any enquiries please email: [Email Address Removed]