Pharmaceutical pollution is a rapidly emerging agent of global change. Psychoactive pharmaceuticals such as antidepressants are of particular concern due to their ability to alter wildlife behaviour at environmentally relevant concentrations. Yet field-based evidence of such effects remains rare. Here, we used novel slow-release chemical implants and high-resolution acoustic telemetry across three experimental lakes to investigate how exposure to a field-realistic mixture of antidepressant pollutants (fluoxetine, sertraline, & venlafaxine) altered daily activity patterns in European perch (Perca fluviatilis; n = 100) and common roach (Rutilus rutilus; n = 79) under natural conditions. Antidepressants accumulated in the brains of exposed fish at concentrations comparable to those reported in wild populations inhabiting polluted waterways. In the smallest, highest-density lake, both perch and roach exposed to antidepressants displayed elevated daytime activity relative to controls, yet no such effects were detected in the two remaining lakes. Despite all three lakes sharing the same habitat types, thermal and light regimes, fish communities, and pollutant concentrations, behavioural effects were lake specific. We surmise that differences in fish density may have partly driven this context dependency. Our findings demonstrate that psychoactive pollutants can alter fish activity under real-world conditions, but highlight that such effects may depend on local ecological context, complicating extrapolation from laboratory experiments to real-world settings.