State-dependent aperiodic EEG dynamics track cortical network reorganization in chronic epilepsy
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Aperiodic (1/f-like) EEG activity has rapidly become a popular noninvasive marker of cortical network state, proposed to index excitation–inhibition (E/I) balance and increasingly applied across neurological and psychiatric disorders. However, whether this approach remains reliable in the pathological brain, where disease progressively reorganizes neural networks, alters signal morphology, and drives continuous transitions between cortical states has yet to be systematically established. Using a medication-free genetic model of chronic epilepsy (Lafora disease; Epm2a−/− mice), we tracked the aperiodic component of the cortical EEG across resting wakefulness, isoflurane anesthesia, and PTZ-induced seizures of graded severity, asking how a single spectral marker behaves as the brain moves between states. Epileptic mice exhibited systematically steeper aperiodic exponents than controls, an effect that persisted after removal of interictal epileptiform discharges and was replicated using independent time-resolved spectral parameterization. Slopes steepened predictably under GABAergic anesthesia, supporting the interpretation that aperiodic activity captures biologically meaningful state transitions beyond simple contamination by pathological waveforms. Across seizure phases, the aperiodic exponent varied systematically, however, the exponent flattened during ictal activity in step with the dominant discharge morphology, revealing that pathological waveform shape itself is a substantial contributor to seizure-state exponent changes. Together, these findings indicate that aperiodic EEG dynamics reflect a combination of chronic network-state reorganization and waveform-shape-driven spectral distortion, with their relative contributions varying across brain states. These results support spectral parameterization as a sensitive approach for tracking pathological neural activity in chronic epilepsy while delineating its interpretive boundaries in the presence of pathological waveforms.
Significance statement
The “aperiodic” structure of brain electrical activity has rapidly become a popular noninvasive marker of cortical state, widely used to infer the balance of excitation and inhibition in health and disease. But how faithfully does it report on a brain that is both diseased and continually changing state? Using a genetic mouse model of chronic epilepsy, we show that this marker conflates two distinct phenomena: lasting reorganization of cortical networks, and distortion caused by the shape of pathological discharges themselves. Separating these contributions across seizure stages and severities yields a more accurate and cautious framework for interpreting this EEG signature; one that matters broadly, as spectral analysis is increasingly applied to noisy, disease- and treatment-altered recordings in neurological and psychiatric patients.
