Interference between cues in human contingency learning: A review, new data and a potential general explanation based on propositional models

Abstract

Interference between cues (IbC) is a phenomenon in which associat- ing a cue with an outcome interferes retroactively with the retrieval of previously acquired association between another cue and the same outcome. In this presentation, we review those situations in which IbC has been found and not found in the literature, and propose a general mechanism responsible of IbC based on propositional mod- els. IbC would take place in those situations in which participants had reasonable expectations of univoque relations between cues and outcomes/responses but later these expectations are not fulfilled. We present three experiments that tested this hypothesis. In the first one, the likelihood of univoque relations was manipulated by using a dif- ferent number of available responses, while in the second partici- pants were pretrained with univoque or non univoque filler relations. In both cases reducing the likelihood of the expectation of univoque relations reduced the magnitude of IbC, supporting the proposed hy- pothesis. Finally, in the third experiment we showed that when the interfering association is presented using instructions it also produces IbC, as in a trial-by-trial learning. Its magnitude was similar or high- er, depending on the type of measurement used. These results suggest that a high-level propositional inference produces IbC.