Critical dynamics of cortical excitability and hippocampal seizures

The sudden emergence of dangerous seizures is the defining feature of epilepsy, but how and when the brain changes dynamics remains enigmatic. From the perspective of dynamical systems theory, such critical transitions occur upon inconspicuous perturbations in highly interconnected systems presenting positive feedback loops and can be modeled as mathematical bifurcations between alternative regimes. Here, we sought to characterize dynamical signatures of impending seizures. We systematically verified the predictions of the model called ‘Epileptor’ that captures seizure dynamics in five differential equations in controlled experiments. We used bidirectional GABAergic pharmacological manipulations to control levels of cortical excitability and gauged resilience to seizures using optogenetic stimulations in healthy mice and direct electrical stimulation in patients with epilepsy. We showed that probing the cortex with single pulses stimulation allows to decode momentary states of cortical excitability (Accuracy: 0.83, p<0.01) and correlate with measurable changes in cortical resilience to seizure (r2=0.20, p<0.01). In mice, we further showed that active probing could anticipate PTZ-induced seizures, acting as a true ‘warning sign’ of ictal transition (+54% 95%CI [36,70] increased response four minutes before transition). Of importance for the design of future neurotechnologies, active probing surpassed passive recording both to decode underlying levels of brain excitability. Our findings provide a promising approach for predicting and preventing seizures, based on a sound understanding of their dynamics.