Clinical trials in sleep disorders report substantial improvement in symptoms in their placebo groups. Behavioral conditioning is one of the underlying mechanisms of the placebo response. However, we do not know whether, and if so, the extent to which sleep architecture is influenced by behavioral conditioning, similarly to other physiological responses (i.e., those in the immune system). We therefore applied a conditioning paradigm to 39 healthy adults pairing a novel‐tasting drink (conditioned stimulus, CS) with the REM sleep suppressing tricyclic antidepressant amitriptyline as unconditioned stimulus during the acquisition phase. Subsequent sole presentation of the CS (together with a placebo pill) in an evocation night led to significantly more REM sleep in the amitriptyline group. Instead of the expected REM sleep suppression in the evocation night, we observed more REM sleep, indicating a rebound that interferes with the conditioned response.