OBJECTIVE This review aims to offer a critical account of recent scientific developments relevant to the hypothesis which Pierre Gloor proposed in the 1970s for the generation of spike and wave discharges (SWDs) of primary generalized absence seizures. RESULTS According to this hypothesis SWDs develop in the same circuits, which normally generate sleep spindles, by an initially cortical transformation of one every two or more spindle waves to a 'spike' component of SWDs, while the next one or more spindle waves are eliminated and replaced by a slow negative wave. This hypothesis was based on experiments in feline generalized penicillin epilepsy showing the possibility of transition from spindles to SWDs, when cortical neurons become hyper-responsive to thalamocortical volleys, which normally induce spindles, and thus engage feedback cortical inhibition, rebound excitation, recurrent intracortical dissemination of excitation during the 'spike' and strong excitation of thalamus for further augmentation of a brain wide synchronous oscillation. In the 1980s, electrophysiological studies in vitro and in vivo revealed the basic features of spindle rhythm generation by neurons in nucleus reticularis thalami and thalamocortical-corticothalamic oscillatory reverberations. CONCLUSIONS In the light of this knowledge, experimental studies in several genetic and pharmacological animal models of absence seizures, clinical observations and theoretical studies in computer models have considered, tested, modified and challenged this hypothesis. It may still be found useful in the era of dynamic digital EEG analysis of SWDs and its current sources.