In the previous paper we have demonstrated, by means of field potential and extracellular unit recordings, that bicuculline-induced seizures, which include spike-wave (SW) or polyspike-wave (PSW) complexes, are initiated intracortically and survive ipsilateral thalamectomy. Here, we used multisite field potential and extracellular recordings to validate the patterns of cortical SW/PSW seizures in chronically implanted, behaving cats. To investigate the cellular patterns and excitability during spontaneously occurring and electrically elicited cortical seizures, we used single and dual intracellular recordings from regular-spiking (RS) and fast-rhythmic-bursting (FRB) cortical neurons, in conjunction with field potential recordings from neocortex and related thalamic nuclei, in cats maintained under ketamine-xylazine anesthesia. 1) Invariably, the spontaneous or electrically induced seizures were initiated within the cortex of both behaving and anesthetized animals. Spontaneously occurring, compound seizures consisting of SW/PSW complexes at 2-4 Hz and fast runs at 10-15 Hz, developed without discontinuity from the slow (mainly 0.5-0.9 Hz), sleeplike, cortically generated oscillation. 2) During SW/PSW complexes, RS neurons discharged spike trains during the depth-negative component of the cortical "spike" component of field potentials and were hyperpolarized during the depth-positive field wave. The FRB neurons fired many more action potentials than RS cells during SW/PSW complexes. Averaged activities triggered by the spiky field potentials or by the steepest slope of depolarization in cortical neurons demonstrated similar relations between intracellular activities and field potentials during sleep and seizure epochs, the latter-being an exaggeration of the depolarizing and hyperpolarizing components of the slow sleep oscillation. 3) During the fast runs, RS cells were tonically depolarized and discharged single action potentials or spike doublets (usually with pronounced spike inactivation), whereas FRB cells discharged rhythmic spike bursts, time locked with the depth-negative field potentials. 4) Neuronal excitability, tested by depolarizing current pulses applied throughout the seizures and compared with pre- and postseizure epochs, showed a decreased number of evoked action potentials during both seizure components (SW/PSW complexes and fast runs), eventually leading to null responses during the postictal depression. 5) Data suggest that interconnected FRB neurons may play an important role in the initiation of cortical seizures. We discuss the similarities between the electrographic patterns described in this study and those found in different forms of clinical seizures.