1. The properties of excitatory postsynaptic potentials (EPSPs) of rat neocortical neurons were investigated with a fast single-electrode current-voltage clamp in vitro. Typically, apparently pure EPSPs were obtained by selection of electric stimuli of low intensity. 2. The amplitude and time integral of the EPSP increased when the neuron was depolarized. At threshold for generation of action potentials, the amplitude of EPSPs was increased by approximately 30% [from 5.0 +/- 2.1 to 6.3 +/- 1.0 (SD) mV, n = 12]. The integral of EPSPs was maximally about fourfold (3.7 +/- 1.5, n = 16) larger than at resting membrane potential (Em). The mechanisms involved in this augmentation of EPSPs were further investigated. 3. The amplitude and the time integral of excitatory postsynaptic currents (EPSCs) decreased linearly with shifts in command potential from -100 to -60 mV. The decrease of the EPSC integral with depolarization indicates that the enhancement of the EPSP may be brought about by recruitment of a voltage-dependent inward current. 4. Evoking EPSPs at various delays after the onset of small depolarizing current pulses (0.3-0.6 nA, 600 ms) revealed that augmentation decays with time. The integral of EPSPs evoked approximately 80 ms after the onset of the current pulse was 3.7 (+/- 1.5, n = 16) times larger than at Em. The integral of EPSPs evoked at 480 ms. however, were only twofold (+/- 0.7, n = 16) larger. Hence EPSPs evoked after a delay of 80 ms were 1.7-fold (+/- 0.4, n = 24) larger than EPSPs evoked after 480 ms. EPSCs were independent of the delay of stimulation at all potentials. 5. Intracellular application of the lidocaine derivative N-(2,6-dimethyl-phenylcarbamoylmethyl) triethylammonium bromide (QX 314) at 100 mM from pipettes rapidly abolished fast action potentials and inward rectification. During comparable depolarizations the increase in EPSP integrals was much smaller in QX 314-treated neurons than in controls. On average, the integral of EPSPs evoked at 70-90 ms was 1.7 times (+/- 1.0) larger than at Em, and the integral of EPSPs evoked with larger delays was close to the value obtained at resting Em (0.9 +/- 0.3, n = 8). The ratio of EPSP integrals early versus late (1.8 +/- 0.5) is comparable to controls, suggesting that QX 314-sensitive currents are unlikely to be involved in the time-dependent enhancement. 6. Mimicking EPSPs by brief depolarizations atop long depolarizations revealed a time- and voltage-dependent enhancement comparable to that of EPSPs.(ABSTRACT TRUNCATED AT 400 WORDS)