Epileptiform activity involves abrupt changes in dynamic behaviour of neuronal ensembles, which alternates between higher complexity ‘interictal’ mode and lower-complexity ‘ictal’ mode characterized by dense, rhythmic firing of the seizing network. Three mechanisms for generating seizurelike events (SLEs) in populations of coupled oscillators will be highlighted as state transitions from higher to lower complexity modes. (i) System parameter changes can cause transitions by way of bifurcations. (ii) Noise fluctuations cause state transitions in bistable systems. This is how paroxysmal transitions are explained in a bistable model of absence epilepsy. (iii) The cognitive rhythm generator network model exhibits intermittency in the absence of either system parameter changes or noise fluctuations. Under simulated epileptogenic conditions, transitions occur unprovoked between the interictal and ictal modes of a chaotic attractor with the trajectory visiting the neighborhood of each mode intermittently. Network "excitability" effects both local and global bifurcations in the dynamics, and under hyperexcitable conditions a bimodal epileptiform attractor is exhibited. This study describes a unified approach to the three mechanisms via a ring device perspective.