Classical fear conditioning involves pairing a neutral conditional stimulus (CS) with an aversive unconditional stimulus (US). Subsequent presentation of the CS alone induces fear responses. Acquisition of conditioned fear is thought to involve learning of the CS/US association, followed by memory consolidation. Recently we reported that the N1/P2 auditory evoked potential was enhanced by fear conditioning in humans. Diazepam 10mg, given before CS/US pairing, prevented subsequent expression of fear potentiation when the response was elicited, 1 week later, in the presence of the CS. In this experiment, we examined whether this effect of diazepam was caused by disruption of the formation of CS/US associations or by disruption of consolidation. The benzodiazepine antagonist flumazenil was used to black the effect of diazepam either during the association period or during subsequent consolidation. Forty-two male volunteers (18-35 years) participated in two sessions separated by 7 days. In Session One, they ingested diazepam 10mg or placebo: 60 minutes later they received flumazenil 1 mg or saline intravenously (i.v.). Then they received 20 presentations of alight (CS), 50% of which terminated with electric shock (US). This was followed by a second infusion of flumazenil or saline. Subjects received placebo/saline/saline (Group 1), diazepam/saline/saline (Group 2), diazepam/flumazenil/saline (Group 3) and diazepam/saline/flumazenil (Group 4). In Session Two, the CS was presented without the US; 50% of CS presentations terminated with a sound pulse; an equal number of sound pulses were presented without the CS. Auditory evoked potentials were recorded at Cz. In Session Two, CS presentation enhanced the auditory N1/P2 potential in placebo-treated subjects (Group 1). This enhancement was prevented by diazepam (Group 2). Flumazenil reversed diazepam's effect on fear potentiation if it was administered before conditioning (Group 3), but not if it was administered afterwards (Group 4). The results confirm that diazepam prevents the acquisition of fear conditioning in humans, and suggest that it disrupts the formation of CS/US associations, rather than the consolidation of fear memory.
