Stimulus property effects on cue competition and temporal estimates during causal learning

Abstract

Learning and timing models have developed along different trajectories within psychology; however, more recent theorising has speculated that both of these phenomena might be modelled within a single theoretical model. While such an approach has merit, the majority of studies into how learning and timing interact have employed nonhuman subjects. Consequently, little is known about how these core psychological processes might interact in humans; this body of experiments was, conducted in order to investigate this issue. Experiments were run to test the hypothesis that cue competition attenuates the ability of participants to estimate a stimulus’ temporal parameters. By studying whether temporal estimates differed between cues in conditions in which blocking and overshadowing was predicted to be weaker or stronger, it could be determined whether time and association were encoded together. In a series of causal learning experiments participants were trained with a cue competition paradigm. On test both cue competition and temporal estimates were examined. The results showed that participant instructions influenced cue competition and that cue properties could influence blocking and overshadowing in specific cases. Temporal estimates made by participants were influenced by cue properties: less accurate estimates of target cue duration were made in several experiments, and temporal estimates between groups varied when blocking and overshadowing were constant. Existing associative learning theories could predict blocking and overshadowing, but could not predict the temporal results. Timing models, for example, the SET model, failed to predict temporal results. To conclude, the results suggest that timing is not encoded as part of the association.