Flexibility of a Conditioned Response: Exploring the Limits of Attentional Capture By Fear

Abstract

Recent work from the attention capture literature suggests that attention may be captured by stimuli with learned aversive value, even when these fear conditioned stimuli (CS) are task-irrelevant and not physically salient. Moreover, relatively little work in the human fear conditioning literature has investigated whether conditioned fear responses can flexibly transfer to a neutral associate of a CS. We examined, for the first time, whether fear-conditioned capture effects were able to transfer to the associate of a CS. Twenty-seven participants encoded novel scene-object pairs. Following encoding, scenes were presented alone during a conditioning phase. Scenes co-terminated with shock 100% (CS100), 50% (CS50), or 0% (CS0) of the time, depending on the object that they had been paired with during encoding, while participants made shock expectancy ratings. Subsequent to conditioning, participants performed a visual search task; the search display occasionally contained one of the encoded objects as a distractor. Eye movements were recorded. Results indicated that, during search, significantly more overt eye movements were made, in error, to the object associate of a CS relative to baseline distractors, and target-directed saccades on trials containing a CS associate were slower relative to target-directed saccades on baseline trials. However, there were no differences in capture effects across the three CS conditions (which varied in threat learning history), suggesting that fear-conditioned capture effects to a CS may not transfer to novel associates encountered for the first time in the episodic context of an experiment.