Dynamic Encoding of Incentive Salience in the Ventral Pallidum: Dependence on the Form of the Reward Cue

Abstract

Some rats are especially prone to attribute incentive salience to a cue (conditioned stimulus, CS) paired with food reward (sign-trackers, STs), but the extent they do so varies as a function of the form of the CS. Other rats respond primarily to the predictive value of a cue (goal-trackers, GTs), regardless of its form. Sign-tracking is associated with greater cue-induced activation of mesolimbic structures than goal-tracking; however, it is unclear how the form of the CS itself influences activity in neural systems involved in incentive salience attribution. Thus, our goal was to determine how different cue modalities affect neural activity in the ventral pallidum (VP), which is known to encode incentive salience attribution, as rats performed a two-CS Pavlovian conditioned approach task in which both a lever-CS and a tone-CS predicted identical food reward. The lever-CS elicited sign-tracking in some rats (STs) and goal-tracking in others (GTs), whereas the tone-CS elicited only goal-tracking in all rats. The lever-CS elicited robust changes in neural activity (sustained tonic increases or decreases in firing) throughout the VP in STs, relative to GTs. These changes were not seen when STs were exposed to the tone-CS, and in GTs there were no differences in firing between the lever-CS and tone-CS. We conclude that neural activity throughout the VP encodes incentive signals and is especially responsive when a cue is of a form that promotes the attribution of incentive salience to it, especially in predisposed individuals.

Keywords: Goal Tracking; Motivation; Pavlovian Conditioning; Rats; Sign Tracking; Ventral Pallidum.

Figure 1.

Every session had 20 lever-CS trials and 20 tone-CS trials in pseudorandom order. In lever-CS trials, a retractable lever was inserted in the cage for 8 s, then removed and followed by a banana pellet (US). Lever-CS trials elicited both sign- and goal-tracking responses, and rats were categorized as STs (n = 12) or GTs (n = 6) based on their behavior during these trials. In tone-CS trials, an 8-s tone preceded the delivery of a banana pellet. Only goal-tracking responses were observed during tone-CS trials, even in rats classified as STs.

Figure 2.

A, During lever-CS trials, STs showed a high probability of lever approach and a low probability of magazine approach. However, during tone-CS trials, STs showed strong goal-tracking behavior and a high probability of magazine approach that did not differ from the lever approach seen in lever-CS trials. B, For STs, the latency to first contact with the lever or magazine did not differ between lever-CS and tone-CS trials. C, Within trials, STs showed similar rates of contact with the lever in lever-CS trials and the magazine in tone-CS trials, both gradually increasing over the time course of the 8-s CS exposure (mean ± SEM, contacts per 0.5-s bins). D–F, GTs showed a high probability of magazine approach in both lever-CS and tone-CS trials (D), no difference in the latency to contact the magazine between lever-CS and tone-CS trials (E), and no difference in the rate of contact with the magazine over the time course of lever-CS and tone-CS trials (F; mean ± SEM, contacts per 0.5-s bins).

Figure 3.

Posterior VP. Sign-tracking to the lever elicited greater neural responses than goal-tracking to either the lever or the tone. Cells were considered responsive to cue exposure if they showed a sustained increase or decrease in firing during the last 7 s of the 8-s period when the CS was present. A, The percentage of cells that responded to cue exposure was significantly greater for the ST-Lever condition compared to the ST-Tone, GT-Lever, and GT-Tone conditions, and this was true for both excitatory and inhibitory responses (*, p < 0.05; **, p < 0.01; ***, p < 0.001). B, In STs, most neurons responded only to the lever-CS and not the tone-CS; however, in GTs, responses were split more equally between the lever, the tone, or both. C, In STs, the magnitude (mean ± SEM) of the cue-induced change in neural activity in responsive cells was greater during lever-CS trials than tone-CS trials. D, In STs, individual responsive cells (lines) showed significantly greater increases or decreases in firing during lever-CS trials compared to tone-CS trials (***, p < 0.001; gray and white bars represent mean firing rate; EX, excitation; IN, inhibition). E, F, For GTs, the magnitude of change in responsive cells (either excitatory or inhibitory) did not differ between lever-CS and tone-CS trials (E) and responsive cells showed no significant differences between lever-CS and tone-CS trials (F). G, Representative cells are shown for a ST (top) and GT (bottom). Blue horizontal dashed lines represent the baseline firing rate, and vertical red lines mark the beginning and end of the 8-s period when the cue was present. The ST cell showed greater cue-induced excitation in lever-CS trials (left) compared to tone-CS trials (right). An example GT cell showed no change during exposure to the lever or tone, although it did show a brief response to cue onset. H, In lever-CS trials, probability difference scores (i.e., the tendency to sign- or goal-track) were significantly correlated with the average change in firing rate for each rat. There was no significant correlation during tone-CS trials. I, No significant group differences were observed in the latency or duration of responses (figures show values for individual units, with median and interquartile ranges).

Figure 4.

Anterior VP. STs showed sustained changes in neural activity during the cue exposure period of lever-CS trials, but these responses were predominantly inhibitory with few excitations. A, The percentage of cells that showed inhibitory responses during cue exposure was significantly greater in the ST-Lever condition than the ST-Tone, GT-Lever, and GT-Tone conditions, although the percentage of excitatory responses was significantly higher for GT-Lever than ST-Lever (**, p < 0.01; ***, p < 0.001). B, Most ST cells were responsive only to the lever and not the tone; however, GT cells responded to the lever only, tone only, or both at similar rates. C, The average magnitude of change (mean ± SEM) is shown for inhibitions in STs, and is much stronger during lever-CS trials compared to tone-CS trials (excitatory responses were too few to graph). D, STs showed significant within-cell differences in inhibition, with lever-CS trials eliciting lower firing rates than tone-CS trials (***, p < 0.001; gray and white bars represent mean firing rate; EX, excitation; IN, inhibition). E, Among the relatively few GT cells that were responsive to cue exposure, there was no difference in the average magnitude of excitatory or inhibitory change. F, Individual GT cells did not show significant changes in firing magnitude between lever-CS and tone-CS trials. G, Representative cells are shown from a ST (top) and a GT (bottom). Blue horizontal dashed lines represent the baseline firing rate, and red vertical lines mark the beginning and end of the 8-s cue exposure period. The ST cell showed a reduction in firing during lever-CS trials that was not seen in tone-CS trials. The GT cell did not show significant changes in firing during exposure to the lever-CS or the tone-CS, which was typical of most GT cells. H, In lever-CS trials, probability difference scores were significantly correlated with average changes in firing for each rat, though there was no correlation in tone-CS trials. I, The characteristics of neural responses (latency and duration) did not differ significantly between groups (figures show values for individual units, with median and interquartile ranges).

Figure 5.

The approximate anatomic location of cells recorded in the VP that showed a sustained response during the 8-s period of cue exposure. Each dot represents a single responsive cell and indicates whether it showed an excitatory response (red) or an inhibitory response (blue). Top row, during lever-CS trials, STs showed a concentration of excitatory responses in the posterior portion of the VP and a concentration of inhibitory responses in the anterior portion of the VP, both of which are visibly diminished in the tone-CS trials. Bottom row, in GTs, both excitatory and inhibitory responses to the CS are sparsely distributed compared to the STs, with little difference between the lever-CS and tone-CS trials.

Figure 6.

VP neurons showed immediate, phasic responses to the onset of lever and tone cues. Top figures represent cells from the posterior VP, and bottom figures represent cells from the anterior VP. A, In the posterior VP, responses to cue onset were largely excitatory. The percentage of cells that were responsive to cue onset did not differ between STs and GTs, although STs had significantly more excitatory cells in lever-CS trials than tone-CS trials (*, p < 0.05). B, Both excitatory and inhibitory responses to CS onset occurred within 0–400 ms and typically returned to baseline in <1 s. C, Responses to CS onset were examined within cells in lever-CS versus tone-CS trials for STs (left) and GTs (right). STs, but not GTs, showed significantly greater excitatory firing in lever-CS trials than tone-CS trials (*, p < 0.05; gray and white bars represent mean firing rate; EX, excitation; IN, inhibition). D, In the anterior VP, there were no significant differences in the proportion of responsive cells. E, F, Firing to CS onset was brief and immediate (E), and for STs (left), but not GTs (right), excitatory responses were significantly greater in lever-CS trials than tone-CS trials (F; **, p < 0.01).

Figure 7.

Many VP cells were responsive to presentation of the food reward (i.e., the US) in both the posterior VP (top row) and the anterior VP (bottom row). A, Although there were no differences between lever-CS trials and tone-CS trials in the posterior VP, STs had a higher proportion of excitatory responses than GTs, and GTs had a higher proportion of inhibitory responses than STs (*, p < 0.05; **, p < 0.01; ***, p < 0.001). B, Responses to the US were prominent during a 1-s period when the food pellet was delivered, which occurred 0.6–1.6 s after the CS phase ended and pellet was released from the food dispenser (shown as time = 8 s). C, In STs (left) there were no significant differences in the magnitude of firing in lever-CS versus tone-CS trials, although GTs (right) showed significantly higher firing rates in lever-CS trials than tone-CS trials (*, p < 0.05; gray and white bars represent mean firing rate; EX, excitation; IN, inhibition). D, In the anterior VP, there were no significant differences in the proportion of cells responsive to the US. E, F, STs (left) had significantly lower firing rates in lever-CS trials than tone-CS trials (***, p < 0.001), whereas GTs (right) showed no differences in firing during lever-CS versus tone-CS trials.

Figure 8.

Firing in the VP was not influenced by movement patterns that differed between sign- and goal-tracking responses. A, Firing rates were determined during periods of active lever interaction or magazine interaction, even if both behaviors occurred during a single trial. B, C, In the anterior VP, there were no differences in the percentage of cells that were responsive during lever interaction or magazine interaction (B), and firing rates of individual cells did not differ significantly between periods of lever interaction versus magazine interaction (C; see Ahrens et al., 2016a for a similar analysis of cells in the posterior VP). For STs, the strength of neural responses was not determined by the intensity of motor activity during sign-tracking as measured by lever presses per trial and latency to contact the lever. D, In the posterior VP, the firing of individual neurons in STs was negatively correlated with lever presses and positively correlated with latency (p < 0.05), indicating slightly weaker neural responses when sign-tracking was more intense. E, In the anterior VP, there were no significant correlations between neural responses and lever presses or latency.