HIGH-RESOLUTION CORTICAL RECORDINGS REVEAL CRITICALITY CHANGES PRECEDING SEIZURE ONSET

Understanding how brain network states evolve before seizures remains a major challenge in epilepsy research and care. Although implantable devices and responsive neurostimulation have improved outcomes for some patients, current systems largely respond after seizure onset and provide limited insight into earlier cortical state changes. Continuous recordings from safe, implantable technologies may help capture gradual shifts in network excitability associated with seizure onset. Cortical excitability and network criticality have been proposed as candidate markers of such changes. We examined how pharmacological modulation of seizure risk alters cortical network dynamics using in vivo recordings from rat hippocampus. Single-unit activity was recorded while seizure risk was increased with a low dose of pentylenetetrazol (PTZ) and reduced with the antiepileptic drug levetiracetam. Network dynamics were quantified using branching ratios, which describe how activity propagates through a network by measuring the average number of downstream activations triggered by a single event. Within minutes of PTZ administration, branching ratios increased significantly from baseline before seizure onset (p < 0.05), consistent with a shift toward a more excitable, seizure-prone network state. In contrast, levetiracetam reduced branching ratios, although this effect did not reach statistical significance. These results support an association between network criticality and seizure risk, and provide quantitative benchmarks for tracking cortical state. Motivated by this framework, we developed Elatus, a high-resolution carbon fibre multi-electrode array enabling stable long-term single-unit recordings with minimal tissue response. Such interfaces may support longitudinal studies of cortical state in epilepsy and inform future implantable monitoring strategies.