1. Synaptic transmission mediated by the N-methyl-D-aspartate (NMDA)-receptor type was studied in neocortex from children undergoing surgical treatment for intractable epilepsy. Intracellular recordings from pyramidal cells were obtained in slices of neocortical tissue by use of microelectrodes. Synaptic responses were induced by electrical stimulation and studied with current-clamp and single-electrode voltage-clamp techniques. The NMDA-receptor-mediated component of the synaptic responses was isolated by addition of 10 microM bicuculline and 30 microM 6-cyano-2,3-dihydroxy-7-nitroquinoxaline (CNQX) in the perfusion solution. 2. In the presence of bicuculline and CNQX, electrical stimulation evoked an excitatory postsynaptic potential (EPSP) in every recorded cell. The amplitude of this EPSP increased when membrane potential was depolarized with injected current. 3. All cells studied in voltage clamp were recorded with microelectrodes containing Cs+ and QX 314. To avoid contamination of the responses from voltage-dependent Ca2+ conductances, membrane potential was held at depolarized potentials until Ca2+ spiking inactivated completely. The evoked excitatory postsynaptic currents (EPSCs) measured at resting membrane potential ranged from 100 to 400 pA. The NMDA receptor-selective antagonist DL-2-amino-5-phosphonopentanoic acid (AP-5) reversibly decreased the current amplitude by 60% for 10 microM and 80% for 30 microM. 4. The current-voltage (I-V) relation showed a region of negative slope conductance between -100 and -20 mV. The largest currents (-250 to -900 pA) were recorded in the range of -45 to -20 mV and reversed between -10 and +10 mV. Removing Mg2+ from the perfusion solution decreased the negativity of the slope, which is consistent with a reduction in the voltage-dependent Mg2+ block of the NMDA-receptor channel. 5. The I-V plots obtained from cells recorded in the most abnormal tissue were averaged and compared with those from the least abnormal tissue. No significant difference was found between these two groups. The averaged plots from the youngest patients (8 and 10 mo old) and those from the oldest (5-15 yr old) patients were also compared, and the results from these two groups were not significantly different.(ABSTRACT TRUNCATED AT 400 WORDS)