Efficacy of oxcarbazepine in treating focal epilepsy based on resting-state EEG functional connectivity and power spectral analyses.

BACKGROUND

Objective electroencephalography (EEG)-based biomarkers are needed to assess oxcarbazepine (OXC) response in patients with focal epilepsy. This study aimed to identify resting-state EEG biomarkers associated with oxcarbazepine efficacy in focal epilepsy using power spectral and functional connectivity analyses.

METHODS

In this retrospective cohort study, 27 drug-naïve patients with focal epilepsy underwent resting-state EEG before treatment and approximately 1 year after initiating OXC monotherapy. Nineteen 10-20 system electrodes were recorded at 256/512 Hz, and preprocessed (resampling, detrending, 50 Hz notch, 0.5-70 Hz band-pass, independent component analysis [ICA] artifact removal). Relative power spectral density (rPSD) was estimated via Welch's method (5 s windows, 50% overlap). Functional connectivity (FC) was quantified by amplitude envelope correlation with correction (AEC-C) within canonical bands, yielding 19×19 matrices. Group comparisons used nonparametric tests with Benjamini-Hochberg false discovery rate (FDR) correction for rPSD and network-based statistic (NBS) for FC (5,000 permutations; initial thresholds t = 2.787, P < 0.01, and t = 3.725, P < 0.001). Clinical outcomes were classified as seizure-free (SF) or not seizure-free (NSF) at 6-24 months after OXC initiation.

RESULTS

Baseline clinical characteristics did not differ between groups. At the on-treatment follow-up EEG, the NSF group exhibited significantly higher θ-band rPSD at multiple frontal-central electrodes (Fp1, Fp2, F3, F4, C3, C4, F7, F8, Fz, Cz) after FDR correction. FC analysis showed stronger δ-band connectivity in the NSF group across frontal-central nodes at P < 0.01; under the stricter threshold (P < 0.001), a robust edge between P3 and F7 remained significantly stronger in the NSF group. No significant between-group differences were observed in other frequency bands after correction.

CONCLUSIONS

These findings suggest that the response to oxcarbazepine in focal epilepsy is linked to differential regulation of slow-frequency brain networks. Persistent low-frequency synchronization reflects ongoing network instability and reduced treatment efficacy, whereas attenuation of pathological slow-wave activity indicates effective network stabilization. These band-limited spectral and network features are promising EEG features associated with treatment response to OXC.