Music as a contextual cue in human predictive learning

In two experiments, we investigated the capacity of musical stimuli in establishing context-dependent behavior using a human predictive learning paradigm. The first experiment employed a conditional discrimination task, in which musical contexts provided information to solve the discrimination problem. During a learning phase, participants were trained to discriminate between two stimuli (X + , Y-) in Context A, with the discrimination being reversed (X-, Y+) in Context B. Results from a subsequent test phase revealed that participants adjusted their responses to the stimuli according to the musical contexts. The second experiment used an ABA renewal design to examine whether changes in musical context could lead to response recovery following extinction. In Phase 1, participants learned to predict an outcome based on Cue X in Context A, but this response was extinguished in Context B during Phase 2. In the Test Phase, returning to Context A led to the recovery of the initially learned response. The present results show that musical stimuli can serve as contextual cues in human learning, and we discuss theoretical as well as clinical implications of our findings. • First demonstration that music can serve as an auditory context for visual stimuli in human predictive learning tasks. • Musical contexts influence behavior in a complex task where stimulus-outcome relationships depend on the musical context. • Changes in musical contexts result in the recovery of previously extinguished responses in an ABA renewal design.