Introduction: About one third of epilepsy patients do not respond to medical treatment, but can still benefit from surgical treatment. Despite recent technological advances in epilepsy surgery, 30%-50% of patients with drug-resistant epilepsy never achieve seizure freedom. This is due, in part, to the lack of objective methods for identification of the epileptogenic brain areas prior to surgical intervention. Several lines of evidence have shown that pathological high frequency oscillations (HFOs) at 100-500 Hz recorded via intracranial EEG are involved in epileptogenesis. However, to date there is no reliable approach to detect and classify pathological from normal HFOs for proper identification of epileptogenic activity.
Methods: We used a graph theoretical analysis of intracranial EEG recordings (subdural grids) in patients with drug-resistant epilepsy in order to identify pathological HFO network activity patterns. The high frequency oscillations were classified into ripple (80-250 Hz), fast ripple (250-500 Hz) and HFO (100-500 Hz) bands. Functional connectivity as mutual information of intracranial EEG signals recorded at each electrode, were quantified and normalized for each pair of electrode time series.
Results: We found that during no-seizure states, cortical networks at all high frequency bands were characterized by stable network structure (modular structure) measured by the average number of nodal communities. Iirregular partitioning of the network architecture led to an increased average number of nodal communities during the pre-ictal period in all high-frequency bands (corrected p = 0.03). This “modular breakdown” was seen on average 3-4 minutes before the electrographic seizure onset.
Conclusions: Functional connectome-based measures of HFO dynamics in contrast to single-channel pathologic HFOs have a high potential in facilitating the development of novel biomarkers for epileptogenesis.
Patient Care: More reliable and clinically practical method for identification of seizure networks can enhance surgical outcome for medication refractory epilepsy.
Learning Objectives: -understand intracranial EEG-based biomarkers of epileptogenic activity
-understand the role of high frequency oscillations in epileptogenesis