The investigational anticonvulsant drug, gabapentin (GP; 1-(aminomethyl) cyclohexaneacetic acid) limited repetitive firing of sodium-dependent action potentials of mouse spinal cord and neocortical neurons in monolayer dissociated cell culture. The effect developed slowly over time with sustained exposure. The IC50 was 1.3 x 10(-4) M for exposure times < or = 60 s, 1.9 x 10(-5) M for 10-60 min, and 4.0 x 10(-6) M for 12-48 h. Hyperpolarization restored sustained firing in the continuing presence of GP. Blockade of action potential firing by GP was frequency (use)-dependent. After preincubation with 2.9 x 10(-5) M GP (5 micrograms/ml), trains of brief stimuli at > or = 50 Hz elicited fewer action potentials than in control solution. Also, at 150 Hz, maximal rate of rise of action potentials decreased progressively with repetitive firing in GP-containing, but not control, solution. After overnight incubation in 2.9 x 10(-5) M GP, the absolute refractory period was prolonged from 2.4 +/- 0.6 ms in control solution (n = 11) to 4.7 +/- 0.3 ms (n = 10; P = 0.02 vs. control), and complete recovery from inactivation was prolonged from 8.0 +/- 1.3 ms to 17.0 +/- 2.6 ms (P < 0.001 vs. control). These findings suggest that GP may alter function of voltage-activated sodium channels, but the mechanism is unproven and may be indirect. Limitation of firing was observed in > or = 50% of neurons at concentrations in the range of those found in plasma and cerebrospinal fluid of patients treated successfully with GP. These results suggest that limiting the rate of firing of sodium-dependent action potentials may contribute to the anticonvulsant efficacy of gabapentin.