Purpose In this experiment we examined whether an unanticipated spatial task could increase the differences between lying and truth telling groups of adolescents. In addition, we explored whether there are some elements of such a spatial task that elicit more diagnostic cues to deception than others. Methods In groups of three, adolescents (N = 150, aged 13–14) either experienced (‘truth tellers’) or imagined (‘liars’) an event. In subsequent individual interviews, the adolescents were asked to provide both a general verbal description of the event (the anticipated task), and a spatial description by making marks on a sketch (the unanticipated task). Next, adults (N = 200) rated the degree of consistency between either the general descriptions or the spatial descriptions from the adolescents in each triad. Results The differences between liars and truth tellers were larger for the spatial markings (the unanticipated task) than for the general verbal descriptions (the anticipated task). Importantly, as predicted, the difference between lying and truth‐telling triads was most manifest for markings of salient (vs. non‐salient) aspects of the event. Conclusions The results suggests that (a) using spatial tasks may be a useful tool for detecting deception in adolescents, but that (b) the assessment of credibility should only draw on the salient aspects of the unanticipated spatial task.