Epileptic networks with motif synchronisation for interictal spike and focus localisation

Epilepsy affects around 65 million people worldwide. It is typically diagnosed and characterised with the use of scalp electroencephalogram. Since seizure frequency is not quite predictable, the main goal of this characterisation is the detection of epilepsy forming spikes between seizures, so called interictal epileptiform discharges. They can indicate the probable brain region where the epilepsy develops, known as epileptogenic focus. The gold standard for this assessment is the time-consuming and laborious task of visual inspection by an epileptologist. However, graph theory and functional connectivity could be used to automate the spatial and/or temporal localisation of the epileptiform events, and help to describe the underlying mechanisms of seizure generation and propagation.

In this talk, I will present the motifs synchronisation method to construct a time varying graph from the whole EEG signal, and its usefulness in localisation of the epileptiforming spikes. I will show how these epilepsy networks could be characterised through clustering analysis, and demonstrate a proof-of-concept results and discuss their clinical relevance.