Stimulus over-selectivity occurs when one of potentially many aspects of the environment comes to control behaviour. In two experiments, adults with no developmental disabilities, were trained and tested in an automated match to samples (MTS) paradigm. In Experiment 1, participants completed two conditions, in one of which the over-selected stimulus was established, and then extinguished by pairing the over-selected stimulus with a novel stimulus and reinforcing the novel stimulus. In the control condition no extinction took place. Experiment 2 replicated this experiment using a between subjects design. In both studies, it was found that when the over-selected cue was extinguished, previously under-selected cues re-emerged to control behaviour in a re-test phase. These findings suggest that over-selectivity is not simply due to an attentional deficit, and that the re-emergence of the under-selected stimuli suggests that the previously under-selected stimuli were learnt about. Moreover, the results suggest that extinction of over-selected cues may have potential applied benefit as an intervention.