Dynamic excitation/inhibition balance preceding seizure onset and its link to functional and structural brain architecture
Duma, G. M.; Cuozzo, S.; Danieli, A.; Hansen, J.; Antoniazzi, L.; Osanni, E.; Vitale, V.; Bonanni, P.; Pellegrino, G.
Show abstract
Altered excitation-inhibition (E/I) balance is a hallmark of epilepsy, yet its dynamic evolution before seizure onset remains unclear. In this study, we investigated time-resolved changes in cortical E/I balance and their network-level and structural correlates in patients with drug-resistant focal epilepsy. High-density EEG recordings encompassing at least one recorded seizure were obtained from pre-surgical evaluations. Using source-reconstructed EEG, we tracked dynamic changes in the aperiodic exponent of the power spectrum--a non-invasive proxy for E/I balance--during preictal and interictal periods. Directed functional connectivity was assessed using spectral Granger causality, and correlations with cortical thickness and neurotransmitter receptor density were evaluated. We found that the aperiodic exponent increased progressively in the minutes preceding seizures, reflecting a widespread shift toward cortical inhibition. This pattern was specific to the preictal state and evident across both epileptogenic and non-epileptogenic regions. Periodic activity in delta and theta bands also increased prior to seizures, supporting a global reorganization of cortical excitability. At the functional profile level, epileptogenic regions demonstrated significantly greater outward connectivity than both their inward connectivity and that of non-epileptogenic areas during the preictal phase. Moreover, E/I balance was differentially related to connectivity patterns: non-epileptogenic regions showed a positive correlation between inhibition (higher aperiodic exponent) and outward connectivity, whereas epileptogenic regions displayed a trend toward the opposite pattern. Structure-function coupling also diverged: cortical thickness positively correlated with inhibitory tone in non-epileptogenic regions but not in epileptogenic ones. Finally, we observed that lower values of muscarinic receptor density corresponded to an E/I balance shifted towards inhibition (ie., higher aperiodic exponent), linking cholinergic tone with preictal E/I dynamics. These findings indicate that seizure onset is preceded by a shift toward inhibition at the whole-brain level, accompanied by distinct patterns of directed connectivity and disrupted structural coupling in epileptogenic networks. This multiscale perspective highlights the interplay between electrophysiology, network topology and cortical architecture systems in the moments leading up to seizure onset. Our results support and enlarge the interictal suppression hypothesis, suggesting a network protective mechanism through inhibition, related to a perturbatory event as the seizure onset. Overall, the dynamic aperiodic exponent emerges as a promising non-invasive marker with potential applications in seizure prediction and network-targeted neuromodulation.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Excitation/Inhibition balance relates to cognitive function and gene expression in Temporal Lobe Epilepsy: an hdEEG assessment with aperiodic exponent 98%
- Graph Theoretical Measures of Fast Ripples Support the Epileptic Network Hypothesis 97%
- Low frequency interictal EEG biomarker for localizing seizures 96%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Drug-resistant focal epilepsy in children is associated with increased modal controllability of the whole brain and epileptogenic regions 96%
- Multimodal in vivo recording using transparent graphene microelectrodes illuminates spatiotemporal seizure dynamics at the microscale 94%
- Distributed harmonic patterns of structure-function dependence orchestrate human consciousness 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.