BACKGROUND: Fear-potentiated startle has been suggested as a translational model for evaluating efficacy of anxiolytic compounds in humans. Several known anxiolytic compounds have been tested as well as several putative anxiolytics. Because results of these studies have been equivocal, the aim of the present study was to examine another pharmacological permutation of the human potentiated startle model by comparing two anxiolytic agents to a non-anxiolytic sedative and placebo. METHODS: Twenty healthy volunteers participated in a double-blind, placebo-controlled, cross-over study with four sessions in which they received single doses of the anxiolytics alprazolam (1 mg) and pregabalin (200 mg), as well as diphenhydramine (50 mg) as a non-anxiolytic sedative control and placebo. The design included a cued shock condition that presumably evokes fear and an unpredictable shock context condition presumably evoking anxiety. RESULTS: None of the treatments reliably reduced either fear- or anxiety-potentiated startle. Alprazolam and diphenhydramine reduced overall baseline startle. Alprazolam was found to only affect contextual anxiety in a statistical significant way after two subjects who failed to show a contextual anxiety effect in the placebo condition were excluded from the analysis. Pregabalin did not significantly affect any of the physiological measures. DISCUSSION: The negative findings from this study are discussed in terms of methodological differences between designs and in variability of startle both between and within study participants. CONCLUSION: Even though fear-potentiated startle may be used to translate preclinical evidence to human populations, methodological issues still hamper the application of this model to early screening of putative anxiolytic drugs.