Temporal-spatial characteristics of phase-amplitude coupling in electrocorticogram for human temporal lobe epilepsy

Abstract

Objective

Neural activity of the epileptic human brain contains low- and high-frequency oscillations in different frequency bands, some of which have been used as reliable biomarkers of the epileptogenic brain areas. However, the relationship between the low- and high-frequency oscillations in different cortical areas during the period from pre-seizure to post-seizure has not been completely clarified. Methods

We recorded electrocorticogram data from the temporal lobe and hippocampus of seven patients with temporal lobe epilepsy. The modulation index based on the Kullback-Leibler distance and the phase-amplitude coupling co-modulogram were adopted to quantify the coupling strength between the phase of low-frequency oscillations (0.2–10 Hz) and the amplitude of high-frequency oscillations (11–400 Hz) in different seizure epochs. The time-varying phase-amplitude modulogram was used to analyze the phase-amplitude coupling pattern during the entire period from pre-seizure to post-seizure in both the left and right temporal lobe and hippocampus. Channels with strong modulation index were compared with the seizure onset channels identified by the neurosurgeons and the resection channels in the clinical surgery. Results

The phase-amplitude coupling strength (modulation index) increased significantly in the mid-seizure epoch and decrease significantly in seizure termination and post-seizure epochs (p < 0.001). The strong phase-amplitude-modulating low- and high-frequency oscillations in the mid-seizure epoch were mainly δ, θ, and α oscillations and γ and ripple oscillations, respectively. The phase-amplitude modulation and strength varied among channels and was asymmetrical in the left and right temporal cortex and hippocampus. The “fall-max” phase-amplitude modulation pattern, i.e., high-frequency amplitudes were largest in the low-frequency phase range [−π, 0], which corresponded to the falling edges of low-frequency oscillations, appeared in the middle period of the seizures at epileptic focus channels. Channels with strong modulation index appeared on the corresponding left or right temporal cortex of surgical resection and overlapped with the clinical resection zones in all patients. Conclusions

The “fall-max” pattern between the phase of low-frequency oscillation and amplitude of high-frequency oscillation that appeared in the middle period of the seizures is a reliable biomarker in epileptogenic cortical areas. The modulation index can be used as a good tool for lateralization and localization for the epileptic focus in patients with epilepsy. Significance

Phase-amplitude coupling can provide meaningful reference for accurate resection of epileptogenic focus and provide insight into the underlying neural dynamics of the epileptic seizure in patients with temporal lobe epilepsy.