Relapse Phenomena from Appetitive-to-Aversive and Aversive-to-Appetitive Counterconditioning within a Human Predictive Learning Task

Abstract

Counterconditioning (CC, Sherrington, 1947) involves training of a single cue with two different outcomes (i.e., cue-O1 then cue-O2) of opposite valence (e.g., appetitive and aversive). Thus, CC can be conducted in either the aversive-to-appetitive (cue-aversive O1 then cue-appetitive O2) or appetitive-to-aversive (cue-appetitive O1 then cue-aversive O2) direction. While CC effectively changes behavior elicited by the cue, this change is highly susceptible to relapse after a retention interval (i.e., spontaneous recovery) or a change of physical context (i.e., renewal). We compared relapse from appetitive-to-aversive and aversive-to-appetitive CC using a human predictive learning task. Participants made wagers of fictitious money based on the presentation of different cues and subsequently gained (appetitive outcome) or lost (aversive outcome) the value of their wager. In Experiment 1, we successfully demonstrated both appetitive-to-aversive and aversive-to-appetitive CC by pairing the target cue (T) with the loss (Loss-Gain condition) or gain (Gain-Loss condition) outcome during the original training phase and later pairing T with the gain (Loss-Gain condition) or loss (Gain-Loss condition) outcome during the CC phase. In Experiment 2, we observed spontaneous recovery from aversive-to-appetitive CC after both 7 day and 5min retention intervals, but recovery from appetitive-to-aversive CC was observed only after the 7day interval. In Experiment 3, we observed renewal only in the aversive-to-appetitive direction. After appetitive-to-aversive CC, responding at test was consistent with CC training regardless of the test context. Therefore, it appears that relapse is more likely from aversive-to-appetitive CC than appetitive-to-aversive CC. We discuss possible explanations to account for this difference.