On 30 July 2025, Prof. Neil Vaughan, Associate Professor of Data Science and AI at University of Exeter, delivered a lecture at City University of Hong Kong titled “AI and DL to Stage and Classify Seizures Using EEG and MRI Neuroimaging Enhancement”.
In his talk, Prof. Vaughan first surveyed the field of AI-driven seizure analysis using electroencephalography (EEG). He reviewed current critical methodologies—including Hellinger distance metrics and particle swarm optimization—for staging and classifying epileptic seizures from open-access EEG datasets. These approaches collectively enhance diagnostic accuracy, enable real-time monitoring, and advance personalized treatment strategies.
Expanding on neuroimaging innovation, he introduced a generative AI framework for correcting motion artifacts in structural and functional MRI. His lab developed PySimPace (for realistic MRI motion simulation) and MoErGAN (a motion artifact simulator), focusing on minimizing algorithmic hallucination while ensuring interpretability and generalizability. The pipeline incorporates visual explainability tools like heatmaps and shaders to enhance clinical confidence. This work also contributes to the MICCAI challenge, aiming to redefine MRI quality standards.
Lastly, Prof. Vaughan discussed translational applications, including a book chapter synthesizing generative AI’s role in medical imaging and ongoing clinical validation. His research bridges computational methods with practical healthcare solutions—from seizure monitoring to artifact-free neuroimaging.
The seminar concluded with an engaged discussion on AI’s expanding frontier in neurological diagnostics, attended by researchers and clinicians at CityUHK’s Department of Neuroscience.
