Lever Insertion as a Salient Stimulus Promoting Insensitivity to Outcome Devaluation

Flexible and efficient decision-making in complex environments can be achieved through constant interactions between the goal-directed and habitual systems. While goal-directed behavior is considered dependent upon Response-Outcome (R-O) associations, habits instead rely on Stimulus-Response (S-R) associations. However, the stimuli that support the S-R association underlying habitual responding in typical instrumental procedures are poorly defined. To resolve this issue, we designed a discrete-trials procedure, in which rats must wait for lever insertion and complete a sequence of five lever presses to obtain a reward (20% sucrose or grain-based pellets). Lever insertion thus constituted an audio-visual stimulus signaling the opportunity for reward. Using sensory-specific satiety-induced devaluation, we found that rats trained with grain-based pellets remained sensitive to outcome devaluation over the course of training with this procedure whereas rats trained with a solution of 20% sucrose rapidly developed habit, and that insensitivity to outcome devaluation in rats trained with sucrose did not result from a bias in general satiety. Importantly, although rats trained with pellets were sensitive to satiety-induced devaluation, their performance was not affected by degradation of instrumental contingency and devaluation by conditioned taste aversion (CTA), suggesting that these rats may also have developed habitual responding. To test whether the discrete-trials procedure biases subjects towards habitual responding, we compared discrete-trials to free-running instrumental responding, and found that rats trained with sucrose in a fixed-ratio 5 (FR5) procedure with continuous presentation of the lever were goal-directed. Together, these results demonstrate that discrete presentations of a stimulus predictive of reward availability promoted the formation of S-R habit in rats trained with liquid sucrose. Further research is necessary to explain inconsistencies in sensitivity to outcome devaluation when rats are trained with grain-based pellets.