Multi-timescale interactions underlying the emergence and evolution of seizure dynamics

Epileptic electroencephalographic (EEG) activity exhibits recurrent seizures, evolving fast and slow wave-forms, state-dependent responses to perturbations, and delayed recruitment between brain regions. How these phenomena emerge from interactions across distinct timescales remains unclear. In this study, we propose a multi-timescale Dual-Component Epileptic EEG (DCE) model that reciprocally couples square-wave bursting and spike-wave subsystems, and we investigate its dynamics using slow-fast and bifurcation analyses, pulse perturbations, and coupled-node simulations. The model generates recurrent seizure-like events that evolve from fast ictal-like activity through spike-wave complexes to slow waves, with spectral evolution qualitatively resembling clinical EEG. Phase-space analysis shows that the epileptogenicity parameter regulates recurrence by shifting an ultraslow nullcline, primarily altering interictal duration. Pulse-induced transitions depend jointly on stimulus amplitude and the ultraslow recovery state. Within individual events, the two coupling directions modulate the fast-burst envelope and drive transitions among oscillatory regimes, respectively. At the coupled-node level, local epileptogenicity and coupling strength organize five dynamical modes. The stimulated node responds on a fast timescale, whereas recruitment of a second node is delayed by ultraslow coupling dynamics, with stronger coupling shortening this delay. The DCE model therefore provides a unified dynamical framework that relates these seizure-related phenomena to mechanisms operating across multiple timescales.