Automatic Detection of Epileptiform Discharges in Extra-cranial Electroencephalography

Aims

• Create a new clinical dataset with improved ecological validity.
• Research the application of machine learning models (classical, ensemble, and deep learning) for seizure detection.
• Use Bayesian optimization for pipeline components and hyperparameters.

Additional Funding

$5,000 Google Cloud Platform Research Credits Program

Contribution

Named grant holder; Data acquisition (NHS Trusts); Data cleaning; Class labelling; Feature engineering; Model design, training, and evaluation; Reporting outcomes.

Technologies Used

High Performance Computing Cluster; Google Cloud