People with more severe hearing loss can be helped to hear again using a cochlear implant - a surgically implanted device that electrically stimulates the cochlear nerve (the nerve of hearing). Part of the device is a set of electrodes within the cochlea (the inner ear). Good speech perception with cochlear implants depends on appropriate adjustment of device parameters, the fitting , by an audiologist. With adults, the initial fitting is usually based on perceptual measurements that require verbal feedback from the patient. When verbal feedback is not possible, the parameters are typically set relative to the minimum level required to get a measurable electrical compound action potential (ECAP) - the summed electrical response from all the nerve fibres that can be recorded by modern cochlear implants. Standard ECAP methods, however, do not lead to a good estimate of perceptual threshold because the rate of stimulation used is much lower than that used for everyday listening; fitting based on ECAPs is therefore suboptimal. Based on our previous animal studies, we are proposing a new method to measure the ECAP threshold in humans. This method involves measuring the variability of an ECAP rather than its average size and will be both more accurate and faster. We will test the approach with cochlear implants patients from Selly Oak Hospital (Birmingham) and the House Ear Institute (Los Angeles); all tests will be in collaboration with Advanced Bionics SARL who will provide the hardware required to test patients.We will combine this method with a computer model to predict the number of cochlear nerve fibres in different regions of the patient's cochlea and determine how the current from each electrode spreads throughout the cochlea. Patient-specific models are required to account for the substantial intersubject variability arising, for example, from underlying pathology, the degree of nerve survival, and electrode placement during surgery. The model will be used to guide fitting and decide, for example, which electrodes should be active. This combination of physiological and computational techniques will overcome the limitation of using ECAPs alone to determine channel interaction (how the nerve activity generated by different electrodes overlap): With the standard use of ECAPs to gauge channel interaction, the effect of fibre distribution and current spread cannot be separated. Such a distinction is clinically important because electrodes need not be made inactive for low nerve survival alone.During the project, these patient-specific models will be extended to predict the pattern of nerve activity in response to more general stimuli. The initial model will be modified so that nerve fibres are simulated by a nerve model we have previously developed. The patient-specific parameters from the model will be selected to match ECAP data, which will require the development of novel physiological methods to improve the reliability of the data. In the later stages of the project, the model will be used to relate perceptual measures to the predicted nerve activity and therefore enable a greater understanding of neural mechanisms underlying the sensory perception of electrical and acoustic stimulation. This will lead to better cochlear implant design. Contemporary fitting by an audiologist is expensive and insufficient to enable a systematic investigation of cochlear implant parameters. In future programmes of work, extended patient-specific models will be used to quickly highlight potentially useful parameter values for standard strategies, e.g. the optimum stimulation rate, and to guide the development of novel strategies. All the above ECAP methods and models will be validated with patients with whom verbal feedback is possible. The enhanced fitting procedures derived in this project are expected to increase speech intelligibility and lead to a better quality-of-life for cochlear implant users.

Nuclear technology is, by definition, based around the principle of subatomic physics and the interaction of radiation particles with materials. Whilst the microscopic behaviour of such systems is well understood, the degree of inhomogeneity involved means that the ability to predict the flux of particles through complex physical environments on the macroscopic (human) scale is a significant challenge. This lies at the heart of how we design, regulate and operate some of the most important technologies for the twenty-first century. This includes building new reactors (fission and fusion), decommissioning old ones, medical radiation therapy, as well as opening the way forward into space technologies through e.g. the development of space-bound mini-reactors for off-world bases and protection for high-tech equipment exposed to high-energy radiation such as satellites and spacesuits. Accurate prediction of how radiation interacts with surrounding matter is based on modelling through the so-called Boltzmann transport equation (BTE). Many of the existing methods used in this field date back decades and rely on principles of simulated (e.g. neutron) particle counting obtained by Monte Carlo and other numerical methods. Input from the mathematical sciences community since the 1980s has been limited. In the meantime, various mathematical theories have since emerged that present the opportunity for entirely new approaches. Together with powerful modern HPC and smarter algorithms, they have the capacity to handle significantly more complex scenarios e.g. time dependence, rare-event sampling and variance reduction as well as multi-physics modelling. This five-year interdisciplinary programme of research will combine modern mathematical methods from probability theory, advanced Monte Carlo methods and inverse problems to develop novel approaches to the theory and application of radiation transport. We will pursue an interactive exploration of foundational, translational and application-driven research; developing predictive models with quantifiable accuracy and software prototypes, ready for real-world implementation in the energy, healthcare and space nuclear industries. This programme grant will unite complementary research groups from mathematics, engineering and medical physics, leading to sustained critical mass in academic knowledge and expertise. Through a diverse team of researchers, we will lead advances in radiation modelling that are disruptive to the current paradigm, ensuring that the UK is at the forefront of the 21st century nuclear industry.

Nuclear technology is, by definition, based around the principle of subatomic physics and the interaction of radiation particles with materials. Whilst the microscopic behaviour of such systems is well understood, the degree of inhomogeneity involved means that the ability to predict the flux of particles through complex physical environments on the macroscopic (human) scale is a significant challenge. This lies at the heart of how we design, regulate and operate some of the most important technologies for the twenty-first century. This includes building new reactors (fission and fusion), decommissioning old ones, medical radiation therapy, as well as opening the way forward into space technologies through e.g. the development of space-bound mini-reactors for off-world bases and protection for high-tech equipment exposed to high-energy radiation such as satellites and spacesuits. Accurate prediction of how radiation interacts with surrounding matter is based on modelling through the so-called Boltzmann transport equation (BTE). Many of the existing methods used in this field date back decades and rely on principles of simulated (e.g. neutron) particle counting obtained by Monte Carlo and other numerical methods. Input from the mathematical sciences community since the 1980s has been limited. In the meantime, various mathematical theories have since emerged that present the opportunity for entirely new approaches. Together with powerful modern HPC and smarter algorithms, they have the capacity to handle significantly more complex scenarios e.g. time dependence, rare-event sampling and variance reduction as well as multi-physics modelling. This five-year interdisciplinary programme of research will combine modern mathematical methods from probability theory, advanced Monte Carlo methods and inverse problems to develop novel approaches to the theory and application of radiation transport. We will pursue an interactive exploration of foundational, translational and application-driven research; developing predictive models with quantifiable accuracy and software prototypes, ready for real-world implementation in the energy, healthcare and space nuclear industries. This programme grant will unite complementary research groups from mathematics, engineering and medical physics, leading to sustained critical mass in academic knowledge and expertise. Through a diverse team of researchers, we will lead advances in radiation modelling that are disruptive to the current paradigm, ensuring that the UK is at the forefront of the 21st century nuclear industry.