EEG signal processing and machine learning for analysis of the responses to deep brain stimulation in epilepsy

Shirani, S ORCID: https://orcid.org/0000-0002-4188-5920, 2024. EEG signal processing and machine learning for analysis of the responses to deep brain stimulation in epilepsy. PhD, Nottingham Trent University.

Preview

Text Thesis_revised_SepehrShirani.pdf - Published version Download (27MB) | Preview

Abstract

Single Pulse Electrical Stimulation (SPES) has emerged as a promising technique for the assessment of drug-resistant epilepsy (DRE). SPES offers a controlled method to probe the excitability and connectivity of neural circuits to identify regions associated with epileptic seizures. Current limited methods employed in practice for processing and diagnosing epilepsy using SPES sometimes fall short of explaining the underlying neurophysiological mechanisms and identifying the regions associated with seizure and seizure onset zone (SOZ). The unique nature of the data recorded during SPES sessions, such as the morphology and inconsistent behaviour of responses even for an individual case in a fixed setup, presents significant challenges for conventional processing algorithms. Such methods may lead to misleading results, underscoring the need for innovative approaches in SPES data analysis.

This PhD thesis focuses on increasing the efficacy of SPES used for the treatment of DRE cases by developing signal processing and machine learning pipelines to investigate the responses to SPES. We aim to answer the critical questions regarding the excitation and inhibition imbalance in regions associated with seizure, the relation between the source of responses to SPES and that of interictal epileptiform discharges (IEDs) as important biomarkers for epilepsy, and also provide accurate and robust tools for identifying the regions responsible for seizure generation focusing on array processing using beamforming.

By leveraging advanced techniques like adaptive and nonlinear signal processing, subspace analysis, single-channel source separation, regularised beamforming, and the related optimisations, this work is committed to improving the accuracy of the diagnosis stage, leading to more effective and individualised treatment strategies for DRE cases.

The excitatory and inhibitory components of brain responses to SPES are separated to study their imbalance associated with seizures. Moreover, more accurate and robust source localisation pipelines, including a distributed beamforming algorithm, are suggested and implemented to identify the regions responsible for generating abnormal responses to SPES. The source of these responses is compared with the origin of other epileptiform activities, such as IEDs and SOZ. The methods developed in this work not only offer a better insight into epilepsy and abnormal activity present in the intracranial electroencephalogram data recorded from DRE cases but also provide more reliable tools for their clinical diagnosis in the future, leading to more efficient and cost-effective DRE treatment techniques.

Item Type:	Thesis
Creators:	Shirani, S.
Contributors:	Name Role NTU ID ORCID Bird, J. Thesis supervisor CMP3BIRDJ https://orcid.org/0000-0002-9858-1231 Mahmud, M. Thesis supervisor CMP3MAHMUM https://orcid.org/0000-0002-2037-8348 Valentin, A. Thesis supervisor SST3VALENA UNSPECIFIED
Date:	August 2024
Rights:	The copyright in this work is held by the author. You may copy up to 5% of this work for private study, or personal, non-commercial research. Any re-use of the information contained within this document should be fully referenced, quoting the author, title, university, degree level and pagination. Queries or requests for any other use, or if a more substantial copy is required, should be directed to the author.
Divisions:	Schools > School of Science and Technology
Record created by:	Laura Borcherds
Date Added:	15 Oct 2025 12:53
Last Modified:	15 Oct 2025 12:53
URI:	https://irep.ntu.ac.uk/id/eprint/54583

Actions (login required)

Edit View

Statistics

Views

Downloads