Aperiodic activity indexes neural hyperexcitability in generalized epilepsy

Generalized epilepsy (GE) encompasses a heterogenous group of hyperexcitability disorders that clinically manifest as seizures. At the whole-brain level, distinct seizure patterns as well as interictal epileptic discharges (IEDs) reflect key signatures of hyperexcitability in magneto- and electroencephalographic (M/EEG) recordings. Moreover, it had been suggested that aperiodic activity, specifically the slope of the 1/ƒ x decay function of the power spectrum, might index neural excitability. However, it remained unclear if hyperexcitability as encountered at the cellular level directly translates to putative large-scale excitability signatures, amenable to M/EEG. In order to test whether the power spectrum is altered in hyperexcitable states, we recorded resting state MEG from male and female GE patients (n = 51; 29 females; 28.82 ± 12.18 years; mean ± SD) and age-matched healthy controls (n = 49; 22 females; 32.10 ± 12.09 years). We parametrized the power spectra using FOOOF to separate oscillatory from aperiodic activity to directly test whether aperiodic activity is systematically altered in GE patients. We further identified IEDs to quantify the temporal dynamics of aperiodic activity around overt epileptic activity. The results demonstrate that aperiodic activity indexes hyperexcitability in GE at the whole-brain level, especially during epochs when no IEDs were present (p = 0.0130, d = 0.52). Upon IEDs, large-scale circuits transiently shifted to a less excitable network state (p = 0.001, d = 0.68). In sum, these results uncover that MEG background activity might index hyperexcitability based on the current brain state and does not rely on the presence of epileptic waveforms. Significance Statement It had long been suspected that electric brain activity is systematically altered in hyperexcitability disorders, such as epilepsy. To date, it remained unclear how pathologic aperiodic activity can be quantified. Kopf et al. demonstrate that aperiodic MEG activity indexes neural hyperexcitability, especially when epileptic discharges were absent; hence, providing a novel non-invasively biomarker that possibly reflects neural excitability at the level of whole-brain recordings.