EEG microstates show different features in focal epilepsy and psychogenic nonepileptic seizures

Abstract Objective Electroencephalography (EEG) microstate analysis seeks to cluster the scalp's electric field into semistable topographical EEG activity maps at different time points. Our study aimed to investigate the features of EEG microstates in subjects with focal epilepsy and psychogenic nonepileptic seizures (PNES). Methods We included 62 adult subjects with focal epilepsy or PNES who received video‐EEG monitoring at the epilepsy monitoring unit. The subjects (mean age = 42.8 ± 21.2 years) were distributed equally between epilepsy and PNES groups. We extracted microstates from a 4.4 ± 1.0‐min, 21‐channel resting‐state EEG. We excluded subjects with interictal epileptiform discharges during resting‐state EEGs. After preprocessing, we derived five main EEG microstates—MS1 to MS5—for the full frequency band (1–30 Hz) and frequency subbands (delta, 1–4 Hz; theta, 4–8 Hz; alpha, 8–12 Hz; beta, 12–30 Hz), using the MATLAB‐based EEGLAB toolkit. Statistical features of microstates (duration, occurrence, contribution, global field power [GFP]) were compared between the groups, using logistic regression corrected for age and sex. Results We detected no differences in microstate parameters in the full frequency band. We found a longer duration (delta: B = −7.680, p = .046; theta: B = −16.200, p = .043) and a higher contribution (delta: B = −7.414, p = .035; theta: B = −7.509, p = .031) of MS4 in lower frequency bands in the epilepsy group. The PNES group showed a higher occurrence of MS5 in the delta subband ( B = 3.283, p = .032). In the theta subband, a higher GFP of MS1 was associated with the PNES group ( B = 5.674, p = .025), whereas a higher GFP of MS2 was associated with the epilepsy group ( B = −6.579, p = .026). Significance Microstate features show differences between patients with focal epilepsy and PNES. EEG microstates could be a promising parameter, helping to understand changes in brain dynamics in subjects with epilepsy, and should be explored as a potential biomarker.