Pathological high frequency oscillations (HFOs) are putative neurophysiological biomarkers of epileptogenic brain tissue. Utilizing HFOs for epilepsy surgery planning offers the promise of improved seizure outcomes for patients with medically refractory epilepsy. This review discusses possible machine learning strategies that can be applied to HFO biomarkers to better identify epileptogenic regions. We discuss the role of HFO rate, and utilizing features such as explicit HFO properties (spectral content, duration, and power) and phase-amplitude coupling for distinguishing pathological HFO (pHFO) events from physiological HFO events. In addition, the review highlights the importance of neuroanatomical localization in machine learning strategies.

Ripple oscillations (80-200 Hz) in the normal hippocampus are involved in memory consolidation during rest and sleep. In the epileptic brain, increased ripple and fast ripple (200-600 Hz) rates serve as a biomarker of epileptogenic brain. We report that both ripples and fast ripples exhibit a preferred phase angle of coupling with the trough-peak (or On-Off) state transition of the sleep slow wave in the hippocampal seizure onset zone (SOZ). Ripples on slow waves in the hippocampal SOZ also had a lower power, greater spectral frequency, and shorter duration than those in the non-SOZ. Slow waves in the mesial temporal lobe modulated the baseline firing rate of excitatory neurons, but did not significantly influence the increased firing rate associated with ripples. In summary, pathological ripples and fast ripples occur preferentially during the On-Off state transition of the slow wave in the epileptogenic hippocampus, and ripples do not require the increased recruitment of excitatory neurons.