BackgroundSide effects play a key role in patients’ failure to take antidepressants. There is evidence that verbal suggestions and informed consent elicit expectations that can in turn trigger the occurrence of side effects. Prior experience or learning mechanisms are also assumed to contribute to the development of side effects, although their role has not been thoroughly investigated. In this study, we examined whether an antidepressant’s side effects can be learned via Pavlovian conditioning.MethodsParticipants (n = 39) were randomly allocated to one of two groups and were exposed to a classical conditioning procedure. During acquisition, 19 participants received amitriptyline and 20 participants received a placebo pill. Pills were taken for four nights together with a novel-tasting drink. After a washout phase, both groups received a placebo pill together with the novel-tasting drink (evocation). Side effects were assessed via the Generic Assessment of Side Effects Scale prior to acquisition (baseline), after acquisition, and after evocation. A score of antidepressant-specific side effects was calculated.ResultsParticipants taking amitriptyline reported significantly more antidepressant-specific side effects after acquisition compared to both baseline and the placebo group. After evocation, participants who underwent the conditioning procedure with amitriptyline reported significantly more antidepressant-specific side effects than those who never received amitriptyline, even though both groups received a placebo.ConclusionsOur results indicate that antidepressant side effects can be learned using a conditioning paradigm and evoked via a placebo pill when applied with the same contextual factors as the verum.