Artificial Intelligence and Stereo-EEG: harnessing semiology to identify the epileptogenic zone in drug resistant focal epilepsy

Approximately 150 000 people are living with focal epilepsy in Australia and over one third have seizures which cannot be controlled with medication. For such patients, surgery to remove the source of seizures (epilepsy surgery), offers the only realistic chance of freedom of seizures and therefore cure. However, not all epilepsy surgery is effective, in part because doctors don’t always accurately identify the true location of where a seizure starts in an individual patient’s brain. The features of a seizure (what a patient initial feels and the sequence of their body movements) can give doctors very useful information where the seizure starts actually starts. However, the way doctors process and use this information is imperfect. I aim to use artificial intelligence software to develop a computer program that takes the features of a seizure and predicts where the seizure starts. This program will be trained with over 50 cases of patients with epilepsy where the source of seizures is definitely known. This program will give doctors a powerful tool to better locate the source of seizures in the brain and help boost the number of patients who can have a cure from epilepsy surgery.