Epileptiform activity induced in rat hippocampal slices by lowering extracellular Ca2+ concentration ([Ca2+]o) was studied with extracellular and intracellular recordings. Perfusing the slices with low Ca2+ (less than or equal to 0.2 mM) or EGTA-containing solutions blocked the synaptic responses of hippocampal pyramidal cells (HPCs). Despite the block, spontaneous paroxysms, termed seizurelike events (SLEs), appeared in the CA1 area and then recurred regularly at a stable frequency. Transient hypoxia accelerated their development and increased their frequency. When [Ca2+]o was raised in a stepwise manner, the SLEs disappeared at 0.3 mM. With extracellular recording from the CA1 stratum pyramidale, a SLE was characterized by a large negative shift in the field potential, which lasted for several seconds. During this period a large population of CA1 neurons discharged intensely and often in synchrony, as concluded from the frequent appearance of population spikes. Synchronization, however, was not a necessary precursor for the development of paroxysmal activity, but seemed to be the end result of massive neuronal excitation. The cellular counterpart of a SLE, as revealed by intracellular recording from HPCs in the discharge zone of the paroxysms, was a long-lasting depolarization shift (LDS) of up to 20 mV. This was accompanied by accelerated firing of the neuron. A prolonged after-hyperpolarization succeeded each LDS and arrested cell firing. Brief (approximately 50 ms) bursts were commonly observed before LDS onset. Single electrical stimuli applied focally to the stratum pyramidale or alveus evoked paroxysms identical to the spontaneous SLEs, provided they surpassed a critical threshold intensity. Subthreshold stimuli elicited only small local responses, whereas stimuli of varied suprathreshold intensities evoked the same maximal SLEs. Thus the buildup of a SLE is an all or nothing or a regenerative process, which mobilizes the majority, if not all, of the local neuronal population. Each SLE was followed by absolute and relative refractory periods during which focal stimulation was, respectively, ineffective and less effective in evoking a maximal SLE. In most slices the spontaneous SLEs commenced at a "focus" located in the CA1a subarea (near the subiculum). SLEs evoked by focal stimulation arose near the stimulating electrode. From their site of origin the paroxysmal discharges spread transversely through the entire CA1 area at a mean velocity of 1.74 mm/s. Consequently, the discharge zone of a SLE could encompass for several seconds the entire CA1 area.(ABSTRACT TRUNCATED AT 400 WORDS)