Stimulus overselectivity and effect of reward history

Abstract

Previous research suggests that stimulus overselectivity – exclusive stimulus control by one or only a few elements of a compound stimulus – occurs because elements of higher salience overshadow elements of lower salience. No research, however, has evaluated why some elements are more salient than others in the absence of a differential on any obvious dimension. Two experiments ran procedures standard in inducing an overselectivity effect. Participants discriminated between two compound stimuli, and then selected between elements of both stimuli in a test phase. However, a preliminary procedure associated differential rewards with different colours, in order to determine whether an element containing colours previously associated with higher magnitudes of reward would be overselected when it was presented as part of a training stimulus. Later, the element expected to be most overselected was put in extinction; any emergence for other elements in a subsequent test phase would indicate that overselectivity for an element would be attributable to reward enhancing the salience of the element rather than enhancing attention towards it. Contrary to predictions, there was no effect of reward history; different elements were overselected by different participants irrespective of the reward previously associated with the colour of each. This meant that relatively underselected elements were put into extinction during that phase. In most instances, emergence was observed for the element put in extinction. Furthermore, emergence for elements was often not accompanied by decreases in selectivity of other elements. These findings suggest that the emergence seen in previous studies may be attributable to an artefact of their procedures rather than evidence of an overshadowing account.