Encoding of Sucrose's Palatability in the Nucleus Accumbens Shell and Its Modulation by Exteroceptive Auditory Cues

Abstract

Although the palatability of sucrose is the primary reason for why it is over consumed, it is not well understood how it is encoded in the nucleus accumbens shell (NAcSh), a brain region involved in reward, feeding, and sensory/motor transformations. Similarly, untouched are issues regarding how an external auditory stimulus affects sucrose palatability and, in the NAcSh, the neuronal correlates of this behavior. To address these questions in behaving rats, we investigated how food-related auditory cues modulate sucrose's palatability. The goals are to determine whether NAcSh neuronal responses would track sucrose's palatability (as measured by the increase in hedonically positive oromotor responses lick rate), sucrose concentration, and how it processes auditory information. Using brief-access tests, we found that sucrose's palatability was enhanced by exteroceptive auditory cues that signal the start and the end of a reward epoch. With only the start cue the rejection of water was accelerated, and the sucrose/water ratio was enhanced, indicating greater palatability. However, the start cue also fragmented licking patterns and decreased caloric intake. In the presence of both start and stop cues, the animals fed continuously and increased their caloric intake. Analysis of the licking microstructure confirmed that auditory cues (either signaling the start alone or start/stop) enhanced sucrose's oromotor-palatability responses. Recordings of extracellular single-unit activity identified several distinct populations of NAcSh responses that tracked either the sucrose palatability responses or the sucrose concentrations by increasing or decreasing their activity. Another neural population fired synchronously with licking and exhibited an enhancement in their coherence with increasing sucrose concentrations. The population of NAcSh's Palatability-related and Lick-Inactive neurons were the most important for decoding sucrose's palatability. Only the Lick-Inactive neurons were phasically activated by both auditory cues and may play a sentinel role monitoring relevant auditory cues to increase caloric intake and sucrose's palatability. In summary, we found that auditory cues that signal the availability of sucrose modulate its palatability and caloric intake in a task dependent-manner and had neural correlates in the NAcSh. These findings show that exteroceptive cues associated with feeding may enhance positive hedonic oromotor-responses elicited by sucrose's palatability.

Keywords: auditory cues; feeding; licking; obesity; sucrose palatability; taste.

Figure 1

Three variants of the brief-access taste tests used in this study. (A) Schematic showing a rat in a behavioral box with a tastant delivery system. The box was equipped with a sipper that delivers, in each lick, a drop of either water or concentrations of sucrose (color-coded). Also seen are two instruments that produce auditory cues: a “click” and a white-noise cue. (B) Schematics of the structure of trials in the Gustatory (upper) and the two Gustatory + auditory (bottom) variants (Start and Start/Stop) of the brief-access tests. The depicted trials are colored as a function of four epochs: Initiation (gray), Reward (blue), Empty licks (white) and Stop licking (cyan). RS indicates the Reward Start, whereas RE shows the Reward End. All trials were classified as being either Complete (where animals lick continuously emitting one single bout) or Incomplete (where they pause licking at least once for an inter-lick interval (ILI) ≥ 0.5 s). In the Gustatory variant, there are no auditory cues, whereas, in the Start task, the Start cue (purple ticks) signals that a new tastant becomes available. In the Start/Stop task there is also a Stop (white noise) cue that commences at RE and continues until the timeout of the Stop licking epoch has finished. (C–E) Representative raster plots from three different animals (upper panels) and PSTHs (bottom panels; all aligned to RS, time = 0 s) of licking responses (red ticks) illustrating the Gustatory (C), the Start (D), and the Start/Stop (E) variants of the brief-access tests. In all cases, trials were sorted as a function of the sucrose concentration (see colored scale at right) and lick bout duration. The horizontal blue lines separate Complete from Incomplete trials. The Inset (dashed rectangle in D) is a zoom from the raster plot of the Incomplete trials in a Start brief-access test, highlighting a large number of trials with one-lick bouts (see arrow). The upper PSTHs was for all (Total) trials, the middle panels for Complete trials, and the bottom panels for Incomplete trials. Note that the licks/s in the Incomplete trials track the sucrose palatability.

Figure 2

Animals exhibit distinct palatability (behavioral) responses that depend on the presence of auditory cue(s). (A) The lick rates (licks/s) for the three brief-access tests aligned to the RS (vertical dashed line) as a function of sucrose concentration. The licking responses are shown for the “Total” trials (upper), for Complete (middle) and Incomplete trials (bottom panels). The n's are the number of trials pooling across animals and sessions. Only the Total and Incomplete trials significantly discriminate among the sucrose concentrations. The arrow in the PSTHs of the Start test highlights the abrupt decrease in lick rate for all tastants after which the lick rate tracks the sucrose concentration. The arrow just after RE in the Start/Stop test indicates that animals rapidly stopped licking. Also shown is the Pearson's correlation (r's) between the lick rate in the Reward epoch and the sucrose concentrations. (B) Cumulative distribution functions (CDFs) for the lick rate (total number of licks during the entire 5 s Reward epoch). (C) The CDFs of the first bout duration over the range of RS for the three tests. The bold horizontal lines at 0.5 on the Y-axis indicate the median bout duration for each sucrose concentration. Note that the first lick bout duration was largest in the Start/Stop test and it was the most positively correlated with sucrose concentrations (see r's). The small arrows in the CDF's indicate the fraction of trials where animals made their first lick bout with a one-lick (bout durations equal 0 s).

Figure 3

Auditory Start/Stop cues increase the Bout index, number of rewards obtained and caloric intake. (A) The Bout indices (the ratio of the number of Complete trials to the number of Complete and Incomplete trials) for the three brief-access tests plotted as a function of sucrose concentration. The order was: Start/Stop (green) > Gustatory (blue) > Start (purple). Values are mean ± sem across trials. (B) The cumulative sum of rewarded (wet) licks in the Reward epoch over the 55 min session for each test. The Start/Stop test had the largest number of rewarded licks. (C) A histogram is displaying the average caloric intake as a function of sucrose concentration for the three brief-access tests. Animals tested in the Start/Stop task consumed significantly more kilocalories (sucrose) than those in the other two brief-access tests. N = 91 sessions for the Gustatory task; 109 for the Start task; 97 for the Start/Stop task. (D) The latency to stop licking the empty sipper after the end of the Reward epoch (RE) for each test. ^*Indicates significant differences with an alpha level of 0.05 in comparison with the Gustatory task in all panels.

Figure 4

Distinct neuronal populations in the NAcSh track either sucrose evoked palatability oromotor responses or the sucrose concentration. (A) A raster plot showing a single unit NAcSh response obtained during a Start/Stop test. The action potentials, depicted as black ticks, are shown as a function of the color-coded sucrose concentrations for Complete and Incomplete trials. Below are the PSTHs of the lick rate and the neuronal activity (Sp/s) for all trials that are aligned to the start of the reward epoch (RS). The vertical white lines in the raster plot and the green rectangle on PSTH depict the time-window where occurs the maximum correlation (Pearson's r) between neuronal activity and Palatability index (i.e., the average lick rate as a function of Trial Type for this session). The bottom panel shows the Palatability index (red) and the number of spikes (black) over the analyzed window with a Pearson correlation (r = 0.67, p < 0.0001). (B) A representative example of a neuronal response with a negative correlation with the Palatability index (r = −0.65, p < 0.0001). Conventions are same as above. Note that briefly before and during the first second of the Reward epoch that this neuronal response was inhibited in a manner independent of the sucrose concentration. (C) A histogram showing the percentage of neurons in each task that significantly correlated with sucrose palatability oromotor responses. The plot shows neurons with negative (r < 0, blue) or positive correlations (r > 0, gray) with sucrose palatability. The n's referring to the actual number of neurons in each subset. Gustatory vs. Start, p = 0.14; Gustatory vs. Start/Stop, p = 0.45; and Start vs. Start/Stop, p = 0.047 (D) Mean ± sem of normalized activity (in Z-score relative to activity in water trials) for all neurons with either positive or negative correlation with the Palatability index (r = 0.3 and r = −0.35, respectively; p's < 0.01). Also shown is the Palatability index (red line, also normalized to water lick rate) and its corresponding correlation coefficients with NAcSh neurons. (E) A raster plot of neuronal responses recorded in a Start test constructed only with Complete trials. In these trials, the licking rate is independent of the sucrose concentration. For visualization purposes, except the last lick in the empty lick epoch (see blue ticks), the ticks for licks were omitted. Vertical white lines in the raster plot and the green rectangle in the PSTH depicts the time-window where the Pearson correlation between neuronal activity and sucrose concentration was maximal. The bottom panel shows the average (± sem) of the neuronal activity during the time-window that maximizes the correlation with the sucrose concentration (r = 0.9, p < 0.001). Also shown is the sucrose-independent lick rate in the same time-window (red line). (F) Representative example of a neuronal response whose firing rate was anti-correlated with the sucrose concentration (r = −0.56, p = 0.0012). It is seen that within the first 1.5 s this neuron fired more for 0% (water) than for 20% sucrose. (G) Histogram showing the percentage of neurons in the Complete trials from the three brief-access tests that significantly correlated with the sucrose concentrations. The n's indicate the number of neurons in each category. Gustatory vs. Start, p = 0.86; Gustatory vs. Start/Stop, p = 0.57; and Start vs. Start/Stop, p = 0.7 (H) The average (± sem) activity of all neurons with sucrose concentration related information that either increase or decrease their activity as the sucrose concentration increased. The average Pearson's r coefficients are also shown for Increases and Decreases (r = 0.43 and r = −0.45, respectively; p's < 0.001).

Figure 5

A population of NAcSh neurons can track sucrose concentrations by increasing their coherence with licking. (A) Upper: A schematic of the temporal sequence of a rat licking a sipper (time = 0 s) and PSTHs of the activity of two lick-coherent neurons aligned to all licks in the 5 s Reward epoch. The upper example is from a neuronal response that was recorded in a Gustatory test whose coherence was similar across Trial Types (see color bar inset; r = 0.03, p > 0.05) and that fired when the animal's tongue made contact with the sipper. Lower: This response shows a lick-coherent neuronal response (recorded in a Start/Stop task) that increased its coherence in a sucrose-concentration dependent manner (r = 0.45, p < 0.05) and that was phased locked to tongue retraction. The gray trace indicates the lick rate (and time between arrows indicates a lick cycle that is each time the animal's tongue make contact with the sipper). (B) The mean (± sem) of the coherence of all 12 neurons that significantly increased their synchrony with licking as a function of sucrose concentration.

Figure 6

Lick-Inactive NAcSh neurons respond phasically to both the Start and Stop auditory cues. (A) The population activity of neurons (Z-score) that was either Active (red) or Inactive (blue) for times around the initiation of licking and during and after the Reward epoch for the Gustatory (upper), Start (middle) and Start/Stop test (bottom panels). Responses to all (Total) licks (licks/s) are shown in gray and the time between peaks indicates one lick cycle (~150 ms) and aligned to the onset of the Initiation epoch (for the Gustatory test) and the Start cue for the Start and Start/Stop tests (left column). For each brief-access test neuronal responses were also realigned to the reward start (RS -middle column) and the reward end (RE (right column). The arrows shown in the Lick-Inactive neurons point to the auditory–evoked phasic activation by both the Start and Stop cues (in Start/Stop and Start tests). Times from −1.5 to −0.5 s before the Initiation epoch were always used as the baseline. (B) Histogram for each test of the percentage of neurons that sustain an Active or Inactive modulation during licking. (C) Histogram showing the percentage of Lick-Active or Lick-Inactive neurons with a significant phasic modulation at the Start cue. (D) The proportion of Lick-Inactive neurons that responded phasically to both the Start (Click) and Stop cues (white noise). Horizontal lines in (C,D) indicate significant differences among tests at an alpha of 0.05 (chi-squared test).

Figure 7

Decoding accuracy of sucrose concentrations/palatability by different NAcSh populations. (A) The temporal decoding dynamics in the three brief-access tests of the five sucrose concentrations (palatability) achieved by the five distinct populations. A sixth group is comprised of neurons that do not belong to any population were named “Not modulated” (gray lines). Values are the average and sem of single trials that were correctly classified around RS (time = 0 s). Thicker lines in all panels indicate significantly above chance level (for five stimuli = 20%, horizontal dashed lines) as ascertained by a one group right-tailed t-test (p < 0.05). (B) The time course of the decoding accuracy of sucrose concentrations/palatability achieved by pooling all groups (all modulated neurons in orange) and the decoding reached by the licking responses per se (black; oromotor palatability). (C) Overall decoding during the Reward epoch (mean ± sem) in the three tests as a function of all the Modulated neurons (orange bars and horizontal dashed lines) and the same group but without the participation of each of the populations. Horizontal green lines are the average levels of decoding achieved by dropping (100 times) a random population of neurons as large as the Palatability population. ^*means significant differences with the Modulated population at the alpha level of 0.05.

Figure 8

NAcSh Palatability-related neurons dynamically track the changes in lick rate elicited by sucrose's palatability over the course of the session. (A) It depicts the population PSTHs of the lick rates in the Reward epoch (Licks/5 s) as a function of Trial Types, divided into blocks of 10th percentile of trials each. The lick rate declined over the course of the session in a concentration dependent manner [Repeated Measures-ANOVA, main effect of Trial Types, F_{(4, 915)} = 86.5, p < 0.0001, effect of time, F_{(4, 9)} = 474.06, p < 0.0001, and significant interaction Trial Types × time, F_{(36, 8235)} = 20.9, p < 0.0001]. (B) Population PSTHs of the firing rate during the Reward epoch (spikes/5 s) of the Palatability-related neurons over the course of the session for each Trial Type. The upper panel depicts neurons that fired more to higher concentrations, whereas lower panel shows neurons with decreasing firing rates as the sucrose concentration increased. Similar results were found by using only the firing rate during the “best-window” of each neuron (see Figure S4).

Figure 9

A population of NAcSh neurons signals the Reward End (RE) by transiently inhibiting their activity after the first available cue signal that the Reward epoch has ended. (A) Raster plots of representative neuronal responses aligned to the RS of neurons recorded in the Gustatory (left), the Start (middle), and Start/Stop (right) tests that, after the reward end (RE), display a transient inhibition. Note that in both Complete and Incomplete trials (separated by a horizontal blue line) the inhibitions in the Start/Stop task were more phase-locked to the RE. (B) The color-coded normalized Z-scores, using a baseline −1.5 to 0.5 s relative to and aligned with RS of all neurons with a phasic inhibition after RE and their corresponding population PSTH below (n = 24, 5, and 23 for Gustatory, Start, and Start/Stop test, respectively). Neurons were characterized as a function of their modulation profile relative to licking: Lick-Active, Lick-Inactive, and Not modulated. Black horizontal lines in the upper panel separate the three groups. On the left side, the vertical bars in color indicate their membership to each subgroup with the same conventions as in Figure 6. The arrow indicates the example shown in rater plot in (A).