Comparison of buspirone with diazepam and fluvoxamine on aversive classical conditioning in humans.

The effects of buspirone, fluvoxamine and diazepam were investigated, using healthy volunteers, in an aversive conditioning paradigm, a putative model for conditioned anxiety. The main prediction was that buspirone, an anxiolytic agent which reduces activity in serotonin (5-hydroxytryptophan, 5-HT) neurones, would attenuate aversively conditioned skin conductance responses. Skin conductance responses were recorded to 10 neutral tones (habituation phase). Tone 11 was immediately followed by a 1-s 90-dB aversive white noise (unconditioned stimulus). The conditioning trial reinstated responding to a second presentation of the tones (extinction phase). Skin conductance response amplitude, inter-response level and spontaneous fluctuations were recorded. There were five treatment groups comprising five men and five women. One control group took placebo, another control group received nothing; there was no effect of placebo on any measure. Diazepam (2 mg, p.o.), a positive comparator, markedly reduced the amplitude of skin conductance responses at all phases of the experiment, but only in women. Buspirone (5 mg, p.o.) had the predicted effect of accelerating extinction but also of unexpectedly accelerated habituation of skin conductance responses. There was a trend to reduce spontaneous fluctuations and no effect on skin conductance level. The effects of buspirone were thus specific to responses to stimuli. Fluvoxamine (25 mg, p.o.) had similar effects to buspirone and diazepam in women. An action common to buspirone, fluvoxamine and diazepam, which may account for their shared effect on conditioned autonomic responses, is the suppression of neural activity in the dorsal raphe nucleus. It is argued that enhanced habituation must involve a different mechanism, such as enhanced 5-HT1A function in the terminal fields of the median raphe nucleus.