The “incentive hope” hypothesis described by Anselme & Güntürkün suggests interesting mechanisms underlying the increase of reward-seeking behavior under uncertainty. We think that this approach could provide a valuable framework for the understanding of a variety of human behaviors that negatively affect well-being, such as overeating, pathological gambling, and substance addiction. Indeed, the uncertainty of the reward delivery seems to be a critical feature of many dysfunctional behaviors. A clear example occurs in the case of pathological gambling, where the uncertainty associated with reward expectations is a key factor in the persistence of the pathological behavior (Linnet et al. Reference Linnet, Mouridsen, Peterson, Møller, Doudet and Gjedde2012). The incentive hope hypothesis thereby represents an interesting research line to pursue to investigate affective processes in humans. On that basis, we formulated some considerations on how the findings from the animal literature described by Anselme & Güntürkün could be translated to a human population.

The first aspect to consider is that the increase in reward-seeking behavior under uncertainty is a phenomenon observed in the realm of Pavlovian conditioning. Decades of behavioral and neuroscience studies in nonhuman animals have led to the proposal that there exists multiple types of behavioral control systems in the brain, among which there is the instrumental system and the Pavlovian system (O'Doherty et al. Reference O'Doherty, Cockburn and Pauli2017). The Pavlovian system allows individuals to assign an outcome's affective value (an outcome could be, e.g., a highly palatable food, and it is typically called an unconditioned stimulus) to a neutral stimulus with which it has been systematically associated (e.g., a metronome sound, typically called conditioned stimulus). Pavlovian conditioning is one of the most fundamental forms of behavior that is widely present across the animal kingdom, and has a deep influence on complex forms of human behavior and cognition (O'Doherty et al. Reference O'Doherty, Cockburn and Pauli2017; Pool et al. Reference Pool, Brosch, Delplanque and Sander2014). The instrumental system allows individuals to learn to perform specific yet arbitrary actions (e.g., pressing lever) that lead to a reward (e.g., food). Within the instrumental system, there is the possibility to select actions in a goal-directed manner, which requires a high level of cognitive resources to represent the states, actions, and available goals and to flexibly compute action plans (Balleine & O'Doherty Reference Balleine and O'Doherty2010). It is important to note that instrumental and Pavlovian associations are governed by distinct mechanisms (Dickinson & Balleine Reference Dickinson and Balleine1994; Ostlund & Balleine Reference Ostlund and Balleine2008; Sennwald et al. Reference Sennwald, Pool and Sander2017). Incentive hope does not seem to involve goal-directed mechanisms, but it rather appears to involve more primary Pavlovian mechanisms. This aspect is particularly relevant given that research investigating human affective processes often targets goal-directed mechanisms (Moors Reference Moors2017). Therefore, experimental research aiming at testing the predictions of the incentive hope hypothesis on a human population should target Pavlovian paradigms rather than paradigms involving goal-directed processes.

The second aspect to consider with respect to potential translational research in humans is that the increase in reward-seeking behavior under uncertainty has been observed in a specific class of Pavlovian responses, namely, sign-tracking behaviors. The importance of the existence of multiple classes of Pavlovian responses has for a long time been underlined in the Pavlovian conditioning literature (Balleine & Killcross Reference Balleine and Killcross2006). There is a classical distinction between two classes of Pavlovian responses: the “preparatory responses” (e.g., heart rate, blood pressure), reflecting the motivational properties of the outcome and the “consummatory responses” (e.g., chewing for a solid food outcome vs. liking for a liquid food outcome Konorski Reference Konorski1967), reflecting the sensory properties of the outcome. It has been suggested that there could be more than two associations between the Pavlovian stimulus and the different aspects of the outcome, for instance, by building associations with the hedonic or temporal aspect of the outcome in addition to the sensory, motivational aspects of the outcome (Delamater Reference Delamater2012; Delamater & Oakeshott Reference Delamater and Oakeshott2007). Each of these associations could elicit a different class of Pavlovian responses within the same individual. The different classes of Pavlovian responses are executed in parallel and are underlain by distinct neuronal networks in both animals and humans (Balleine & O'Doherty Reference Balleine and O'Doherty2010; Zhang et al. Reference Zhang, Mano, Ganesh, Robbins and Seymour2016). This suggests that Pavlovian learning is not a unitary process but rather involves a multitude of stimulus outcome associations triggering a multitude of Pavlovian responses. Crucially, these parallel Pavlovian responses can react differently to motivational manipulations: Some classes of Pavlovian responses might be more sensitive to changes in outcome values than others. Therefore, the class of Pavlovian responses measured while testing the effect of uncertainty on human reward-seeking behaviors appears to be fundamental. This highlights the relevance of using Pavlovian paradigms measuring sign-tracking behaviors on a human population.

Nonetheless, to date little research has investigated sign- as opposed to goal-tracking behaviors in humans (Garofalo & di Pellegrino Reference Garofalo and di Pellegrino2015). In animals, sign- and goal-tracking behaviors are distinguished during presentation of the Pavlovian stimulus (e.g., a lever cue that has been associated with the delivery of food in a cup): Some animals approach and engage with the Pavlovian stimulus itself (e.g., approach the lever cue, sign-tracking), whereas other individuals approach the location where the food will be delivered (e.g., approach the food cup, goal-tracking). These inter-individual differences in Pavlovian conditioning are thought to reflect different learning styles that can predict impulsive behaviors and might represent a vulnerability to some mental disorders such as substance abuse (Joyner et al. Reference Joyner, Gearhardt and Flagel2018). Moreover, a consistent amount of research has shown that humans are also attracted to the Pavlovian stimuli themselves: For instance, human attention is very rapidly oriented toward the Pavlovian stimuli (Bucker & Theeuwes Reference Bucker and Theeuwes2017; Pool et al. Reference Pool, Brosch, Delplanque and Sander2014). However, despite the potential contribution that sign- versus goal-tracking propensities could have for identifying risk profiles for substance addiction, the development of paradigms allowing measurement of these two classes of Pavlovian responses in humans is just beginning (Joyner et al. Reference Joyner, Gearhardt and Flagel2018). The development of such paradigms will be extremely valuable to research trying to translate the incentive hope” hypothesis from animals to humans.

In conclusion, we suggest that research testing the prediction of the incentive hope hypothesis in humans should aim at using Pavlovian paradigms and measuring a specific class of Pavlovian responses consisting of sign-tracking behaviors. Such a translational effort of the incentive hope mechanisms could provide valuable information for the understanding of several dysfunctional behaviors observed in humans such as overeating and gambling.