Differential inflation with short and long CS-US intervals: Evidence of a nonassociative process in long-delay taste avoidance

Three experiments were performed to investigate the learning process underlying the phenomenon of long-delay taste conditioning. An associative model views taste avoidance as due to a conditioned stimulus-unconditioned stimulus (CS-US) associative structure, despite the long interval interposed between the flavor and illness. A nonassociative account of this avoidance behavior posits that avoidance stems from the interaction of two nonassociative processes: habituation of neophobic avoidance to a novel taste, and the poison-induced dishabituation of this process. A postconditioning inflation manipulation was used to discriminate between these two views. It has been demonstrated that enhanced responding with a US inflation manipulation depends, in part, on a previously conditioned association. Therefore, if long-delay taste avoidance arises from nonassociative processes, an inflation manipulation should not affect conditional responding. Experiment 1 demonstrated a delay of reinforcement effect, enhanced avoidance of saccharin in the immediate/inflation group, and no effect of inflation in the delay group and sham controls. Experiment 2 revealed that this differential effect of inflation is not due to absolute differences in the strength of the avoidance response. In Experiment 3, we investigated a potential associative learning mechanism that could account for the differential inflation effect. Together, the present results support the various predictions of a nonassociative account of long-delay taste avoidance.