In this project we studied how the functional connectome derived from EEG recordings can be used to localize the epileptic focus in patients. The functional connectome indicates how brain regions communicate with each other. For example in the figure we see a functional connectome during a seizure. The arrows show the connections between the regions during the seizure. Based on these connections we can see that the brain region depicted by the star is sending information to other regions in the brain during the seizure. This indicates that this brain region is leading the seizure and we can consider it as the epileptic focus.
We investigated the functional connectome in 27 patients during 111 seizures. We compared the standard EEG-based localization technique, namely looking which brain region is most active during a seizure, with the localization technique based on functional connectome that depicts the driver of the seizure. All patients had epilepsy surgery after the presurgical evaluation and were seizure-free after surgery. This allowed us to compare the brain region we localized based on the two techniques with the resection in the patients. The brain region with maximal activity was inside the resection in 31% of the seizures and estimated within 10 mm from the border of the resection in 42%. Using the functional connectome, these numbers increased to 72% in the resection and 94% within 10 mm of the resection. Therefore, we showed that looking at the functional connectome during seizures has an added value and should be included in the presurgical evaluation.
We also investigated the functional connectome during resting state in 20 patients with left temporal lobe epilepsy, 20 patients with right temporal lobe epilepsy and 35 healthy age-matched subjects. We studied how well we can classify a person to have epilepsy or not, and if we can predict the lateralization of the epilepsy (left vs. right). The diagnosis and lateralization classifiers achieved a high accuracy (90.7% and 90.0% respectively). Meaning that based on 15min resting state EEG we can predict if the person has epilepsy or not, and if yes, the side where the epilepsy is originating from with 90% accuracy.
The project resulted in a total of 8 peer-reviewed publications: 2 in NeuroImage Clinical, 2 in Brain Topography, 1 in Epilepsia Open, 1 in Brain Stimulation, and 1 in IEEE Transactions on Biomedical Engineering and over 20 conference contributions. The results of the project have been orally presented at 10 conferences such as the European Congress on Epileptology 2016 & 2018, the International Epilepsy Congress 2017 and the International Congress of Clinical Neurophysiology 2018. The first prize of oral presentation was won at the Alpine Brain Imaging Meeting 2016.
