Backward Sensory Preconditioning when Reinforcement is Delayed

Abstract

In two experiments rats received an initial phase of training in which two neutral stimuli were presented as a serial compound (A-X). In a second phase, A was established as a signal for a shock reinforcer, the shock being presented immediately after the termination of A (the immediate condition) or after a 5-sec interval (the trace condition). A final test phase showed that not only A but also X was capable of evoking conditioned suppression (a backward sensory preconditioning effect). The degree of suppression evoked by X was not correlated with that evoked by A. In both experiments the A trained with immediate reinforcement was more suppressive than that trained with the trace procedure, but in Experiment 1 the trace and immediate groups did not differ in the response they showed to X, and in Experiment 2 (which allowed a within-subject comparison) the trace procedure resulted in more suppression to X than did immediate conditioning of A. These results disconfirm the suggestion that the backward sensory preconditioning effect depends on the formation of an associative chain: X-A-shock. They are consistent with the proposal that the associatively activated representation of X is able to form a direct association with the reinforcer during A-shock training.