How Learning and Motivational Shifts Determine the Incentive Salience of Reward Cues: a Human Behavioral Perspective

Reward cues acquire distinctive incentive properties from their association with motivationally significant outcomes. These incentive properties are manifest in an augmented salience that provides reward cues with the capacity to attract attention, also in contexts where pursuing the reward is not relevant. I will first examine the unresolved debate concerning the learning parameters that define the cue-reward association and eventually modulate attention. Specifically, whether the cue attentional salience is governed by reward expectancy or uncertainty. I will then study another incentive property of reward cues, that is the ability to strengthen the performance of a separately learned instrumental action exerted to obtain an outcome, a phenomenon known as Pavlovian-Instrumental transfer (PIT). Since the motivational value of reward is altered when reward is devalued, an interesting question is whether the incentive properties of reward cues change accordingly, both in the attentional and in the operant domain. Therefore, I will investigate the effects of motivational shifts on the incentive properties of reward cues, by means of reward devaluation. In the context of the ability of a cue attentional salience to update in accordance with the altered outcome value, I will further analyze the influence of incentive learning (i.e. the possibility to re-associate the cue with the devalued outcome). The procedures adopted through the experiments share two main phases: a learning phase in which human thirsty participants learned cue-beverage reward associations involving different contingencies, and a test phase in which no reward was delivered. During the learning phase, participants accumulated the beverage reward that acted as an incentive, since it was not consumed through this phase. This allowed a controlled devaluation procedure by consummatory satiation (i.e. a motivational shift) that was administered at different moments during the experiments depending of the hypothesis at test. Results showed that the cues that better predicted the reward during learning were the stimuli preferentially attended at test, and the stimuli that invigorated more the instrumental action for the outcome. These incentive attributes persisted despite reward devaluation: the attentional bias and the PIT effect emerged unaltered after participants quenched their thirst. Reward cues persisted in capturing attention after reward devaluation even when participants were given the chance for incentive learning by means of an additional learning phase. Taken together, the evidence that emerged indicates that the incentive properties of reward cues, once acquired, can surprisingly and irrationally outlast reward devaluation and can resist incentive learning, suggesting that some incentive properties of the cue can become independent from those of the reward. These results may provide important implications for the understanding of the psychological mechanisms underlying different types of addiction.