My current research interests lie in biological time-series analysis and control engineering with a particular focus on epileptic seizure prediction.
Epilepsy is a neurological disorder characterised by recurrent "seizures" and is associated with abnormal neuronal activity in the brain. Epilepsy the single most common serious neurological condition in the world today, with the World Health Organisation estimating that the prevalence of epilepsy varies across geographical regions within the range of 0.5% to 4% of the total population. The ability to predict seizures would have a profound impact on the quality of life of epilepsy suffers. An ideal solution would be the development of an implanted device, incorporating seizure prediction and an electric stimulation treatment, which could prevent a seizure from occurring once an imminent seizure is predicted. While electric stimulation treatment for epilepsy is already undergoing trial, no robust seizure prediction algorithm has been published to date. With this motivation for seizure prediction, this research aims to investigate a suitable algorithm for prediction of the onset of epileptic seizures. It is proposed that entropy estimates as computed from models based on data compression techniques shall be investigated particularly models based on the Willems weighted context tree algorithm, Markov processes and non-linear Independent component analysis (ICA). The data to be used with these models is the electroencephalogram which results from recordings of the fluctuating electric fields of the brain.
This research in epileptic seizure prediction also aims to investigate a theoretical framework for dynamical system time-series analysis that is not based on delay-reconstruction/ embedding. This framework is to be developed particularly with the application of real-world biological data in mind that is, noisy, non-stationary data from a high dimensional underlying dynamical system.