Computer modelling, software engineering, and AI to predict epilepsy surgery success using medical imaging

Epilepsy is the recurrence of spontaneous unprovoked seizures which often involve a loss of consciousness and abnormal neural activity. Epilepsy is often associated structural and functional brain abnormalities detectable by MRI and EEG/MEG. Surgical treatment for epilepsy aims to remove the part of the brain thought to be causing seizures. However around 40% of patients will continue to experience seizures even after such an invasive operation. It is currently not possible to know which patients will be rendered seizure-free by surgery, and which will not before the operation.

Aim

The aim of this PhD is to develop predictive computational models which can be used for 1) refining suggested surgical resection strategies and 2) predicting patient outcomes for a given strategy.

Methods

In this PhD we will use MRI data from patients with epilepsy to infer a personalised brain network. We will then use the network to constrain parameters in a computer model of neural dynamics. The model dynamics will be fit to the patient’s own neural dynamics, measured by MEG. After meeting this initial objective of model fitting, we will modify the parameters to perform simulated brain surgery to investigate if the model ‘seizes’ for a given simulated surgery. Model outputs will be compared against patient outcomes for validation. We have data acquired from over 100 patients who already underwent surgery and where the outcome is known.

Timeliness

This work would be the first of its kind to combine MRI and MEG data using a dynamical model in epilepsy, with the aim to ultimately contribute to improved patient outcomes. The work is timely considering the use of AI and advanced modelling, alongside improved accessibility to high performance computing.

Potential impact

Through our extensive collaborator network we have the opportunity to move towards clinical translation, influencing clinical decision making within the lifetime of the PhD. This opportunity has the potential to improve our mechanistic understanding of seizures, and the optimal way to treat them.

Supervisory team

We have a wide range of expertise in computing, medical imaging, and statistics in the lead supervisory team. We already collaborate extensively with colleagues in the medical faculty, and will continue in this manner. Dr Ahmadi, has extensive expertise in signal processing and medical computing, with an interest in application to epilepsy. 

Funding

PhD studentships are funded by the Reece Foundation for 4 years. Funding will cover tuition fees at the UK rate only, a Research Training and Support Grant and a stipend (Year 1: £22,000, Year 2: £23,000. Year 3: £24,000. Year 4: £25,000). Applications are welcomed from students in all countries, although students from outside the UK will be required to pay full international fees. International students may be eligible for additional financial support to cover some, or all, of these fees.

Enquiries

Professor Peter N Taylor [Email Address Removed]

Centre for Neuroscience: [Email Address Removed]

Applications

https://www.ncl.ac.uk/research/transformative-neuroscience/studentship/