The influence of instructions on reversing the generalization of valence, US expectancy, and electrodermal responding in fear conditioning.

Pairing a conditional stimulus (CS) with an aversive unconditional stimulus (US) causes negative valence and US expectancy to generalize to stimuli that are perceptually and/or conceptually similar to the CS. Past research has shown that instructing participants that the US is more likely to follow stimuli that are dissimilar to the CS reversed the generalization of US expectancy but left generalized valence unchanged. Here, we examined whether instructions about the relationship between stimuli that are perceptually similar would affect the generalization of valence. A picture of an alien (CS+) was paired with an electric stimulus, while a perceptually different alien stimulus (CS–) was presented alone. After conditioning, valence, US expectancy, and electrodermal responses generalized to different aliens that were perceptually similar (by color and shape) to the CS+ and CS–. Participants were then instructed that aliens perceptually similar to the CS+ belonged to the same group as the CS‐ and that aliens perceptually similar to the CS– belonged to the same group as the CS+. The instructions caused an elimination (but not a reversal) of generalized expectancy and valence but did not affect generalized electrodermal responses. This suggests that evaluations of generalization stimuli are sensitive to instructions about their relationship to the CS and that dissociations reported in the literature between valence and expectancy after instructions may occur due to the type of instruction used. Past research has shown that instructions about the likelihood of the unconditional stimulus occurrence affect unconditional stimulus expectancy ratings during perceptually similar generalization stimuli but do not influence evaluations. In the current study, we demonstrate that instructions that change how the conditional and generalization stimuli are related affected generalized unconditional stimulus expectancy ratings and evaluations but did not affect electrodermal responding. [ABSTRACT FROM AUTHOR]