Neural mass modeling reveals that hyperexcitability underpins slow-wave sleep changes in children with epilepsy

Abstract

Objective: The relationship between sleep and epilepsy is important but imperfectly understood. We sought to understand the mechanisms that explain the differences in sleep homeostasis observed in children with epilepsy. Methods: We used a neural mass model to replicate sleep electroencephalography (EEG) recorded from 15 children with focal lesional epilepsies and 16 healthy age-matched controls. Different parameter sets were recovered in the model for each subject. Results: The model revealed that sleep EEG differences are driven by enhanced firing rates in the neuronal populations of patients, which arise predominantly due to enhanced excitatory synaptic currents. These differences were more marked in patients who had seizures within 72 h after the sleep recording. Furthermore, model parameters inferred from patients resided closer to model parameters inferred from a typical seizure rhythm. Significance: These results demonstrate that brain mechanisms relating to epilepsy manifest in the interictal EEG in slow-wave sleep, and that EEG recorded from patients can be mapped to synaptic deficits that may explain their predisposition to seizures. Neural mass models inferred from sleep EEG data have the potential to generate new biomarkers to predict seizure occurrence and inform treatment decisions.