The incentive salience of the stimuli biases rats’ preferences in the “suboptimal choice” procedure

Highlights

• We presented to rats asymmetrical arrangements of incentive salience in the alternatives of a “suboptimal choice” procedure.

• We found a strong bias for choosing the alternative with higher incentive salience.

• The bias for choosing the alternative with higher incentive salience was stronger for sign-tracking subjects.

Abstract

In the suboptimal choice procedure, pigeons systematically choose an alternative associated with stimuli that indicate the presence or absence of a reward and with a smaller probability of reinforcement, over another option associated with noninformative stimuli, and a higher probability of reinforcement. In contrast, rats generally show the opposite preference. A recent hypothesis for explaining this difference suggests that both species would prefer the discriminative alternative in situations in which the stimuli have incentive salience. However, this hypothesis has not been confirmed in rats. The null relationship between incentive salience and performance in the suboptimal choice procedure stands in contrast to the strong influence that incentive salience has been shown to have in other choice protocols. A possible explanation for this difference is that in the suboptimal choice procedure, the incentive salience is normally shared by both alternatives and subjects attribute the same incentive salience to them. To analyze this possibility, we presented an asymmetrical arrangement of the incentive salience between the discriminative and the non-discriminative alternatives in order to evaluate its impact on suboptimal choice. We also examined whether rats classified as sign-trackers are more influenced by this asymmetrical arrangement. The results showed that the alternative with higher incentive salience attracted more preference and that this effect was stronger for sign-tracking subjects. These results suggest a strong relationship between preference in the suboptimal choice procedure and the incentive salience attributed to the different alternatives.

Introduction

In recent years, the suboptimal choice procedure has been extensively employed for studying the mechanisms of choice in humans (Molet et al., 2012), pigeons (Stagner and Zentall, 2010), rats (Trujano and Orduna, 2015) and other species (Vasconcelos et al., 2015), probably because of its relationship with some instances of maladaptive behavior in humans, like pathological gambling (Molet et al., 2012). In the suboptimal choice procedure, two alternatives are presented: discriminative, and non-discriminative. When the first is chosen, two different stimuli may be presented, one that signals with certainty that a reinforcer will be delivered or another that signals -also with certainty- that no reinforcer will be available. When the second alternative is chosen, either of two stimuli may be presented, both associated with the same probability of reinforcement, which is usually smaller than 1.0 -making the reinforcer uncertain- but more likely than that of the discriminative alternative, which makes it the optimal alternative. The procedure was named after pigeons showed a strong preference for the discriminative alternative despite it being associated with a lower overall probability of reinforcement (for reviews, see Zentall, 2014; McDevitt et al., 2016).

One of the unresolved issues regarding this procedure is the notable difference that has been found when comparing the performance of pigeons and rats: as stated above, in at least dozens of experiments pigeons have shown a strong and consistent preference for the discriminative alternative (Kendall, 1974; Smith and Zentall, 2016; Stagner and Zentall, 2010). In contrast, accumulating evidence shows that rats prefer the optimal alternative, even though they readily discriminate the different stimuli associated with it (Alba et al., 2018; López et al., 2018; Martinez et al., 2017; Ojeda et al., 2018; Trujano and Orduna, 2015; Trujano et al., 2016). Although different possibilities have been suggested to account for this species difference (Cunningham and Shahan, 2018, 2019; Chow et al., 2017; Trujano et al., 2016; Zentall et al., 2019), currently there is no agreement on its cause.

One of the recent hypotheses regarding this difference lies on the concept of incentive salience (Chow et al., 2017). This hypothesis is based on the fact that for pigeons, the discriminative stimuli employed in the suboptimal choice procedure are presented on keys, and it has been shown that illuminated keys evoke sign-tracking behavior in pigeons (Brown and Jenkins, 1968), which has been taken as evidence that pigeons attribute incentive salience to them. In contrast, the first experiments on suboptimal choice with rats presented lights as stimuli, which evoke goal-tracking behavior in rats (Beckmann and Chow, 2015; Boakes, 1977), suggesting that lights do not have incentive salience for them.

In this context, Chow et al. (2017) evaluated rats in a suboptimal choice procedure in which the discriminative stimuli were levers for one of the groups, and lights for a comparison group; the result was that the “levers” group showed suboptimal behavior while the “lights” group showed optimal behavior. Unfortunately, because in this study the negative outcome of the discriminative alternative was not associated with a lever, but with a blackout, it was possible that the suboptimality found by Chow et al., was due to the absence of a conditioned inhibitor, a stimulus that has been shown to impact rats’ behavior in this procedure (Trujano et al., 2016). To analyze this possibility, Martinez et al. (2017) performed an experiment in which one lever was associated with the positive outcome of the discriminative alternative and another lever with the negative one. Two additional levers were associated with the non-discriminative alternative. Under these circumstances, rats strongly preferred the non-discriminative (optimal) alternative, even though the discriminative stimuli had incentive salience.

A stronger test of the hypothesis that relates suboptimal choice with attribution of incentive salience was performed by Lopez et al. (2018); in this study subjects were classified by their capacity of attribution of incentive salience using the Pavlovian Conditioned Approach procedure, under the hypothesis that subjects with high capacity of attribution of incentive salience would show a higher level of suboptimal behavior than subjects with less capacity. However, the results indicated a high preference for the optimal alternative in both groups, suggesting either that incentive salience does not play a role in the suboptimal choice procedure, or that the influence of incentive salience was cancelled because it was shared by both alternatives. Other studies have also failed to replicate the result by Chow et al. (2017), despite the use of levers as discriminative stimuli (Alba et al., 2018; Orduña and Alba, 2019). In marked contrast to these results, in a related research area it has been shown that the incentive salience of the stimuli has a strong effect on preference. Beckmann and Chow (2015) associated in different trials the presentation of a lever or a tone with pellet delivery; after this phase, subjects were presented with a choice between these stimuli and showed a strong preference for the lever (95 % preference). This preference remained (75 % preference) even when the probability of pellet delivery in presence of the lever diminished to 0.50.

It is possible that the difference between Beckman and Chow's (2015) study and those that have reported no relationship between suboptimal choice and the incentive salience of the stimuli (Alba et al., 2018; Orduña and Alba, 2019; Martinez et al., 2017; López et al., 2018) is due to the fact that in the former study the alternatives differed in the incentive salience of their associated stimuli, while in the second group of studies the incentive salience was the same in both alternatives.

In the present study, we manipulated the incentive salience of the discriminative and the non-discriminative alternatives by employing levers in one of them, and lights in the other, in order to evaluate its impact on suboptimal choice (experiment 1). In addition, this manipulation was replicated employing samples of subjects with high and low capacity of attribution of incentive salience (experiment 2), under the hypothesis that the influence of the incentive salience would be higher in the former.