The unpredictability of epileptic seizures is perhaps the most disabling aspect of this common condition. We recently performed a first-in-man study of long-term monitoring of intracranial EEG for the purpose of seizure prediction, demonstrating that real-time seizure prediction is feasible, and can lead to new approaches to the management of epilepsy.
The study also permitted collection of continuous EEG in the ambulatory setting, providing new insights into the patterns of seizure activity.
We have discovered complex and highly individual relationships between spikes and seizures, and unexpected circadian, ultradian, and infradian rhythms. Applying sophisticated analysis techniques to this data has allowed identification of the dynamics underlying the patterns of seizure activity, potentially providing new therapeutic opportunities.
We have extended this work recently in partnership with IBM, showing that advanced chip technologies combined with novel deep-learning strategies can further improve seizure prediction.
We are currently commercialising a less invasive implantable seizure monitoring systems in collaboration with partners at the Bionics Institute and St. Vincent’s Hospital. Ultimately our ambitions to combine predictive methods with delivery of therapeutics, through both electrical counter stimulation techniques and advanced drug delivery systems.
- Baldassano SN, Brinkmann BH, Ung H, Blevins T, Conrad EC, Leyde K, Cook MJ, Khambhati AN, Wagenaar JB, Worrell GA, Litt B.
- Crowdsourcing seizure detection: algorithm development and validation on human implanted device recordings. Brain. 2017 Jun 1;140(6):1680-1691. doi: 10.1093/brain/awx098. PubMed PMID: 28459961