Although sodium channels have been reported to be inactive after 5-10 minutes of ventricular fibrillation (VF), their state during early VF is unknown. In 12 open-chest dogs, a floating glass microelectrode was used to record intracellular action potentials from the right ventricle during pacing and during electrically induced VF. Before any drug was administered, an initial episode of VF was continuously recorded for at least 20 seconds followed by defibrillation. Recordings were made during VF episodes after superfusion for 15 minutes around the microelectrode site by low (2.8x10-5 M) and high (10-4 M) concentrations of tetrodotoxin (TTX) in five dogs, or by low (4 μM) and high (100 μM) concentrations of verapamil in another four dogs. In three dogs, VF was induced without drugs three times to determine if the effects observed in the previous dogs were caused by the drugs or by successive episodes of VF. Ten consecutive action potentials were analyzed at the onset and after 5, 10, 15, and 20 seconds of VF. Action potential amplitude and duration during paced rhythm or VF were not changed by the local perfusion of either TTX or verapamil. In the TTX group, the maximum upstroke rate of depolarization of an action potential (V̇(max)) during paced rhythm was 104±14 V/sec for control cycles before any drug was given, 86±15 V/sec for the low TTX concentration, and 55±14 V/sec for the high TTX concentration (p<0.05 versus other two). V̇(max) decreased from 55±32 V/sec at the beginning of VF to 37±27 V/sec after 20 seconds of VF for predrug VF, from 39±20 V/sec to 18±11 V/sec for low-dose TTX VF, and from 18±13 V/sec to 12±7 V/sec for high-dose TTX VF (p<0.05 among the three groups). In the dogs receiving verapamil, VF was still inducible with V̇(max) not significantly different from predrug VF at the onset and after 5 or 20 seconds of VF but with V̇(max) smaller (p<0.05) for verapamil than for predrug VF after 10 or 15 seconds of VF. In three dogs, V̇(max) was not significantly different during three successive episodes of VF when no drug was given between the episodes. Thus, the ability of TTX to decrease V̇(max) indicates that some sodium channel activity remains in early VF and the inability of verapamil or TTX to prevent VF suggests that both fast and slow channel activity maintains VF during the first 20 seconds.