Quantifying individual variation in the propensity to attribute incentive salience to reward cues

Abstract

If reward-associated cues acquire the properties of incentive stimuli they can come to powerfully control behavior, and potentially promote maladaptive behavior. Pavlovian incentive stimuli are defined as stimuli that have three fundamental properties: they are attractive, they are themselves desired, and they can spur instrumental actions. We have found, however, that there is considerable individual variation in the extent to which animals attribute Pavlovian incentive motivational properties ("incentive salience") to reward cues. The purpose of this paper was to develop criteria for identifying and classifying individuals based on their propensity to attribute incentive salience to reward cues. To do this, we conducted a meta-analysis of a large sample of rats (N = 1,878) subjected to a classic Pavlovian conditioning procedure. We then used the propensity of animals to approach a cue predictive of reward (one index of the extent to which the cue was attributed with incentive salience), to characterize two behavioral phenotypes in this population: animals that approached the cue ("sign-trackers") vs. others that approached the location of reward delivery ("goal-trackers"). This variation in Pavlovian approach behavior predicted other behavioral indices of the propensity to attribute incentive salience to reward cues. Thus, the procedures reported here should be useful for making comparisons across studies and for assessing individual variation in incentive salience attribution in small samples of the population, or even for classifying single animals.

Figure 1. Individual variation in the propensity to approach the lever-CS or food cup after 5 days of Pavlovian training.

The number of lever deflections (A) and food cup entries (B) is shown for 1,878 individual rats (Subject Number), with the order of the cases randomized so the values are not clustered. Note that there is enormous individual variation in the preferred response.

Figure 2. The distribution of each of the three components of the PCA Score across each of the 5 days of Pavlovian training.

Panel A shows the Response Bias score, Panel B the Probability Difference score and Panel C the Latency score. The number of rats with a given PCA Score are binned into 20 bins of equal size (0.1 bin sizes), according to their score, which ranges from +1 to −1. Thus, the vertical axis shows the number of rats in each bin, and the horizontal axis the PCA Score. Note that the Response Bias score and the Probability Difference score show the development of two subpopulations by Days 4 and 5 of training.

Figure 3. The distribution of PCA Scores across each of the 5 days of Pavlovian training, using the formula given in Table 1.

The number of rats are binned according to their PCA Scores, which ranges from +1 to −1, with 0.1 bin sizes. The PCA Scores range from +1 to −1. Thus, the vertical axis shows the number of rats in each bin, and the horizontal axis the PCA Score. Note that PCA Score reveals two subpopulations of animals by Days 4 and 5 of training.

Figure 4. The PCA Score is strongly correlated with lever- and food cup-directed behavior on the final day of training.

Selected scatterplots of correlations reported in Table 2 are shown. Each symbol represents an individual animal. The top panels show the number of lever contacts plotted as a function of the PCA Score on Day 1 (A) and Day 5 (B) of Pavlovian training. The bottom panels show the number of head entries into the food cup, during the 8 s CS period (during which time the lever was inserted into the chamber), plotted as a function of the PCA Score on Day 1 (C) and Day 5 (D) of training.

Figure 5. Variation in the topography of the conditioned response from trial to trial as a function of PCA Index Score.

This analysis is based on a subset (n = 370) of the total sample of animals. Behavioral responses on a given trial were classed as: (1) ONLY trials (a trial in which a rat made only one or more lever deflections or only one or more food cup entries during the CS period, but not both); (2) BOTH trials (a trial in which an animal made at least one lever deflection and one food cup entry in the same 8-s CS period); (3) NONE trials (trials in which there was neither a lever deflection nor a food cup entry). Panels A and B show the percent of ONLY and BOTH trials for each rat, respectively, plotted as a function of the animal's PCA Index Score. Based on these data we classed the animals as sign-trackers (STs; a PCA Index Score of 0.5 or above), goal-trackers (GTs; a score of −0.5 or less), or intermediates (INs; scores from −0.49 to +0.49). Panel C shows the proportion of ONLY, BOTH, and NONE trials for STs, INs, and GTs. Panel D shows a more detailed analysis of the Intermediates. The INs are subdivided into those with positive vs. negative PCA Index Scores. This does not have much effect of the percent of ONLY, BOTH or NONE trials, but the inset shows that INs with positive scores (towards sign-tracking) typically press the lever on ONLY trials, and those with negative PCA Index Scores typically make food cup entries.

Figure 6. The distribution of PCA Index Scores in a sample of 1,878 rats. STs, GTs and INs are classed according to the criteria presented in Fig. 5.

It is clear, based on these criteria, that STs and GTs represent different subpopulations of animals.

Figure 7. Mean ± SEM number of lever deflections (A) or food cup entries (B), probability of approaching the lever (C) or food cup (D) during the CS period, and latency to contact the lever (E) or make a food cup entry (F) during the CS period, over 5 days of Pavlovian training in 1,878 rats classed as STs, GTs or INs, as described in Fig. 6.

Note that the SEM is smaller than the symbol in most cases.

Figure 8. Animals in which the CS and US are presented but not paired do not learn either a ST or GT CR.

(A) The mean (± SEM) number of lever presses in STs and the Unpaired group (UN) and (B) CS food cup entries in GTs and the Unpaired group. The data are reanalyzed from Robinson & Flagel , and STs and GTs were re-classed based on the PCA Index.

Figure 9. Mean (SEMs are occluded by the symbols) frequency of food cup entries (entries/s) during the CS period vs. intertrial (ITI) periods in sign-trackers (STs) and goal-trackers (GTs) over five days of Pavlovian training.

The STs and GTs are those shown in Fig. 6 and 7. The data illustrate that GTs discriminate CS vs. non-CS periods, selectively increasing head entries during the CS period as a function of Pavlovian training.

Figure 10. The propensity to approach a lever-CS predicts the ability of the same lever-CS to support learning a new instrumental response to get it (i.e., the ability of the lever-CS to act as a conditioned reinforcer).

Data from Lomanowska et al. were used to compare the effectiveness of the rank-order split and PCA Index methods to predict the ability of the CS to act as a conditioned reinforcer. For the rank-order split method, rats were classed as STs and GTs by totaling the number of lever contacts over 5 days of Pavlovian training and dividing the sample of animals tested into thirds. Panel A shows the correlation between active nose-pokes (minus inactive nose-pokes) on the test for conditioned reinforcement, as a function of total lever contacts. Panel B shows the same data, but when each animal's PCA Index Score was calculated and used to class animals. In both Panels red filled symbols indicate GTs, white symbols INs, and blue filled symbols STs, classed the two different ways. Horizontal lines depict group means. (Note that the sample sizes differ for the groups between the two methods; an equal number of STs and GTs cannot be assumed when using the PCA Index.).