Article Figures & Data
Figures
- Figure 1.
Ventral Pallidal neurons are cue-responsive and their firing reflects cue value. A, Illustration of the behavioral task. Lp, lever press; rwd, reward; cue, DS or NS. B, Raster plot (upper panels) showing firing aligned to DS and NS onset for a representative neuron. Movement onset is depicted for each trial by a green dot. Histogram (lower panel) shows mean firing rate across all correct trials for DS (red) and NS (black). C, Heatmaps (upper panels) depict DS and NS response for all cue-excited neurons (N = 165). Color represents Z-score of firing rate as compared with a 1-s precue baseline. The lower panel shows the mean frequency of firing rate for neurons significantly excited by the DS. D, Mean firing response to DS aligned to DS onset for first-quartile (shortest) movement onset latency (upper plot) and fourth-quartile (longest) movement onset latency trials (lower plot). Red line depicts average cue-evoked firing on DS trials in 98 DS-excited neurons. Blue and green lines depict cumulative percentage of locomotor onsets and lever presses, respectively. Note that the firing response to DS onset is similar in the two latency conditions, indicating that little to none of the excitation is because of a secondary excitation at movement onset. E, Same as C but firing is aligned to movement onset for all trials. F, Correlation of normalized (Z-score) response to the NS versus DS (40- to 180 ms window following cue-onset) for each neuron. Shapes indicate the animal from which the neuron was recorded. Red line indicates best-fit linear regression (r2 = 0.52). Slope <1 indicates that neurons respond more strongly to DS than NS.
- Figure 2.
Scree plot of factor analysis on locomotor variables and cross-correlation coefficients of variables selected for GLM. A, A scatter plot showing the normalized magnitude of the scalar component of each principal component (eigenvalue), where the mean value of all eigenvalues is 1. Because five components have values >1, we considered these five factors in subsequent FA of the data. B, Correlation matrix showing the correlation coefficients for each pair of variables used in the GLM analysis.
- Figure 3.
Cue-evoked firing of VP neurons was correlated with movement speed toward the lever and with lever proximity at cue onset. Analysis results from the larger (10-variable) GLM are shown. A, Each bar indicates mean and SE of normalized regression estimates (IDR Firing Difference, the percent change in firing response across the IDR of the variable) for 165 cue-excited neurons in the overall firing response window (40–400 ms after DS onset). Dark shaded bars indicate significantly correlated regressors. B, Individual distributions of three regressors across cue-excited neurons. Gray bars indicate individually significant coefficients. C, E, Same as A but for the early and late response windows, respectively. D, F, Same as B but for the early and late response windows, respectively. In B, D, F, the vertical bar indicates the mean, and the color of the bar indicates whether the mean is significantly different from 0 (red, p < 0.05, one sample t test with Holm’s modified Bonferroni correction).
- Figure 4.
Representative neurons with firing related to distance from the lever at cue onset. Each raster/histogram pair shows one neuron’s firing in selected DS trials aligned to DS onset. Rasters show the time at which the neuron fires and are sorted from trials where the animal is closest to the lever (top) to farthest (bottom). Histograms show the mean firing rate across trials.
- Figure 5.
Comparison of IDR Firing Differences across rats. Each box-and-whisker plot describes an individual rat’s neurons’ IDR Firing Difference values for the indicated variable. The heavy horizontal line is the median, the box delimits the interquartile range, the lower whisker shows the smaller of the 25th percentile minus 1.5 times the interquartile range or the lower limit of the data range; the upper whisker shows the larger of the 75th percentile plus 1.5 times the interquartile range or the upper limit of the data range; and the individual points show values outside the whisker range; p values are the result of cross-subject ANOVAs.
- Figure 6.
Focused GLM. Analysis results from the focused (three-variable) GLM are shown. Figure panels are as described for Figure 3.
- Figure 7.
Bilateral but not unilateral infusion of the D1 antagonist SCH-23390 into the NAc severely impairs behavioral performance. Box plots show the mean (horizontal line), interquartile range (box), extremes (whisker), and outliers (points) of the response ratio for the indicated cue (DS or NS) preinfusion and postinfusion of the indicated drug (for detailed description of box-and-whisker plots, see legend to Fig. 5). Unilat., unilateral.
- Figure 8.
The D1 antagonist SCH-23390 infused into the NAc does not significantly alter VP neuronal responses to DS onset, except when infused bilaterally. A–E, Heat maps depicting firing responses of neurons that were significantly activated by DS onset. Data from the same neurons (in the same order) are shown before (pre, upper) and after (post, lower) infusion. F–J, Correlation of mean Z-score of response in the 40- to 180-ms window following DS-onset before and after infusion. K–O, Same as F–J but for the response window of 180–400 ms following cue-onset.
- Figure 9.
Recording electrode (X) and infusion cannula sites (solid circles) were localized following completion of experimental session from 20-μm sections. Coordinates were determined by comparison with nearby landmark features (Paxinos and Watson, 1998). Numbers indicate anteroposterior distance of the coronal section from bregma (mm).
Tables
Variable Units Locomotor Variables 1 Radial velocity with respect to operant lever (mean) mm/s 2 Radial velocity with respect to operant lever (maximum) mm/s 3 Radial velocity with respect to operant lever (SD) mm/s 4 Speed (mean) mm/s 5 Speed (SD) mm/s 6 Speed (maximum) mm/s 7 Move duration s 8 Path length mm 9 Latency to time of maximum speed s 10 Latency to time of maximum acceleration s 11 Maximum distance between actual path and best (straight line) path mm 12 Turn efficiency (cumulative summed change in heading/net change in heading) None 13 Angular velocity with respect to operant lever (SD) Radians/s 14 Angular velocity with respect to operant lever (maximum) Radians/s 15 Angular velocity with respect to operant lever (mean) Radians/s 16 Signed change in head orientation Radians 17 Radial velocity with respect to operant lever (mean) mm/s 18 Path efficiency None Variables measured at moment of cue onset 19 The distance from the lever at movement-onset mm 20 Time elapsed since last reward received s 21 Time elapsed since last lever press s 22 Time elapsed since last cue (DS or NS) s 23 Head orientation with respect to lever Radians A total of 23 locomotor and behavioral variables were used for GLM analysis.
Subject Sessions Response
ratio (DS)Response
ratio (NS)DS, locomotor
latency (s)NS, locomotor
latency (s)DS, mean
locomotor
speed (m/s)NS, mean
locomotor
speed (m/s)μ σ μ σ μ σ μ σ 1 7 0.93 0.11 1.16 1.66 2.10 2.07 0.34 0.13 0.29 0.13 2 11 0.92 0.10 1.39 1.66 2.24 2.10 0.36 0.15 0.33 0.14 3 5 0.91 0.13 1.06 1.41 2.82 2.60 0.26 0.11 0.23 0.08 4 2 0.90 0.06 1.27 1.49 2.62 2.42 0.37 0.13 0.31 0.11 Grand mean 0.92 0.10 1.22 1.56 2.45 2.30 0.33 0.13 0.29 0.12 Data from four subjects were used for analysis in this study. This table shows the number of sessions from each subject that contributed data for analysis in this study as well as the response ratios, locomotor onset latency, and mean speed of approach to the operandum for each subject across all sessions for both DS and NS responses (defined as pressing the active lever operandum within 10 s of cue presentation).
Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Rep score Maximum of radial velocity (with respect to operandum) 0.94 0.25 0.16 0.01 −0.03 0.55 SD of radial velocity (with respect to operandum) 0.92 0.14 0.13 0.30 0.00 0.36 SD of speed 0.88 0.10 0.37 0.22 0.00 0.18 Maximum speed 0.88 0.24 0.39 −0.02 0.02 0.25 Move duration 0.09 0.95 0.06 −0.07 0.08 0.78 Path length 0.34 0.82 0.29 0.33 0.05 −0.18 Latency to maximum speed 0.04 0.77 0.09 −0.08 −0.02 0.73 Latency to maximum acceleration 0.04 0.75 0.08 −0.09 0.00 0.72 Maximum deviation from direct path 0.21 0.66 0.28 0.58 0.10 −0.50 SD of angular velocity (with respect to operandum) 0.39 0.12 0.87 0.17 0.00 0.19 Maximum of angular velocity (with respect to operandum) 0.45 0.23 0.82 −0.02 0.19 −0.02 Mean of angular velocity (with respect to operandum) 0.18 0.23 0.74 0.45 −0.38 0.26 Turn efficiency 0.01 0.04 0.04 −0.20 0.02 NA Mean radial velocity (with respect to operandum) 0.54 0.20 0.14 0.77 0.20 −0.30 Mean speed 0.49 0.24 0.37 0.75 0.01 −0.36 Path efficiency −0.03 −0.02 0.11 0.48 −0.02 0.43 Movement onset latency −0.12 0.01 −0.04 −0.15 0.01 NA Net change in heading −0.01 −0.03 0.00 −0.01 0.01 NA PCA followed by FA was used as the basis to select a subset of variables for subsequent use as regressors in GLMs to detect correlation with neuronal activity. Variables used as regressors are shown in bold. We selected a representative variable for each factor based on high loading onto its representative factor and low co-loading onto other factors (which together are represented by the rep score, the variable’s highest factor loading minus the sum of its loadings onto all other factors). These representative variables were: factor 1, maximum of radial velocity with respect to operandum; factor 2, move duration; factor 3: SD of angular velocity with respect to operandum; factor 4, path efficiency. Additionally, we selected three “independent” variables that did not load strongly onto any one factor (turn efficiency, net heading change, and movement onset latency). NA, not applicable.
Comparison Figure Test Statistic p value Multiple comparison procedure DS-evoked firing vs NS-evoked firing 1F Wilcoxon V = 18013 p < 0.0001 None DS-evoked firing: behavioral response vs no response n/a Wilcoxon V = 11007 p < 0.0001 None Large model, overall firing window tests for IDR firing difference from 0 Maximum radial velocity 3A,B t 2.53 p = 0.13 Holm Move duration 3A t 1.33 p = 0.19 Holm SD of angular velocity 3A t 0.43 p = 0.67 Holm Path efficiency 3A,B t 2.61 p = 0.10 Holm Net Δ in heading 3A t 1.67 p = 0.10 Holm Turn efficiency 3A t 0.50 p = 0.62 Holm Latency to move 3A t −1.40 p = 0.17 Holm Distance from lever 3A,B t −1.37 p = 0.17 Holm Time since reward 3A t −0.96 p = 0.34 Holm Time since cue 3A t 1.60 p = 0.11 Holm Large model, early firing window tests for IDR firing difference from 0 Maximum radial velocity 3C,D t 1.75 p = 0.084 Holm Move duration 3C t 1.95 p = 0.054 Holm SD of angular velocity 3C t 0.92 p = 0.36 Holm Path efficiency 3C,D t 3.48 p = 0.00075 Holm Net Δ in heading 3C t 1.56 p = 0.12 Holm Turn efficiency 3C t −0.76 p = 0.45 Holm Latency to move 3C t −1.52 p = 0.13 Holm Distance from lever 3C,D t 0.65 p = 0.52 Holm Time since reward 3C t −0.35 p = 0.73 Holm Time since cue 3C t 0.71 p = 0.48 Holm Large model, late firing window tests for IDR firing difference from 0 Maximum radial velocity 3E,F t 3.48 p = 0.00075 Holm Move duration 3E t 1.00 p = 0.32 Holm SD of angular velocity 3E t 0.64 p = 0.52 Holm Path efficiency 3E,F t 0.75 p = 0.45 Holm Net Δ in heading 3E t 1.38 p = 0.17 Holm Turn efficiency 3E t 2.74 p = 0.0074 Holm Latency to move 3E t 0.27 p = 0.79 Holm Distance from lever 3E,F t −3.66 p = 0.00042 Holm Time since reward 3A t 0.62 p = 0.54 Holm Time since cue 3A t 1.27 p = 0.21 Holm Comparison of IDR firing response across rats Path efficiency, early excitation window 5A ANOVA F = 0.8727 p = 0.458 Maximum radial velocity, late excitation window 5B ANOVA F = 0.32 p = 0.812 Lever distance, late excitation window 5C ANOVA F = 6.84 p = 0.0003 Focused model, overall firing window tests for IDR firing difference from 0 Maximum radial velocity 6A,B t 2.18 p = 0.032 Holm Path efficiency 6A,B t 2.22 p = 0.029 Holm Distance from lever 6A,B t −1.64 p = 0.10 Holm Focused model, early firing window tests for IDR firing difference from 0 Maximum radial velocity 6C,D t 1.24 p = 0.22 Holm Path efficiency 6C,D t 3.26 p = 0.0015 Holm Distance from lever 6C,D t 0.46 p = 0.65 Holm Focused model, late firing window tests for IDR firing difference from 0 Maximum radial velocity 6E,F t 2.42 p = 0.017 Holm Path efficiency 6E,F t 0.69 p = 0.49 Holm Distance from lever 6E,F t −4.33 p = 0.000037 Holm Regression of Q2 vs Q1 early firing Z scores for no infusion Intercept 8F t 0.62 p = 0.54 None Slope 8F t 9.89 p = 3.59 × 10−13 None Slope vs 1 8F t −1.03 p = 0.31 None Regression of Q2 vs Q1 early firing Z scores for bilateral vehicle Intercept 8G t 0.36 p = 0.72 None Slope 8G t 4.50 p = 0.00022 None Slope vs 1 8G t −0.07 p = 0.945 None Regression of Q2 vs Q1 early firing Z scores for contralateral SCH-23390 Intercept 8H t 0.108 p = 0.915 None Slope 8H t 5.92 p = 4.87e-6 None Slope vs 1 8H t 0.35 p = 0.73 None Regression of Q2 vs Q1 early firing Z scores for ipsilateral SCH-23390 Intercept 8I t t = 1.23 p = 0.23 None Slope 8I t t = 5.81 p = 2.68e-6 None Slope vs 1 8I t −1.12 p = 0.27 None Regression of Q2 vs Q1 early firing Z scores for bilateral SCH-23390 Intercept 8J t −0.04 p = 0.97 None Slope 8J t 3.79 p = 0.0016 None Slope vs 1 8J t −8.37 p = 3.09e-7 None Regression of Q2 vs Q1 late firing Z scores for no infusion Intercept 8K t 0.28 p = 0.78 None Slope 8K t 10.71 p = 2.53e-14 None Slope vs 1 8K t −1.25 p = 0.22 None Regression of Q2 vs Q1 late firing Z scores for bilateral vehicle Intercept 8L t 0.79 p = 0.44 None Slope 8L t 8.44 p = 5.08e-8 None Slope vs 1 8L t −0.18 p = 0.857 None Regression of Q2 vs Q1 late firing Z scores for contralateral SCH-23390 Intercept 8M t 0.51 p = 0.61 None Slope 8M t 6.41 p = 1.53e-6 None Slope vs 1 8M t −2.36 p = 0.027 None Regression of Q2 vs Q1 late firing Z scores for ipsilateral SCH-23390 Intercept 8N t 0.80 p = 0.43 None Slope 8N t 8.16 p = 5.37e-9 None Slope vs 1 8N t −1.10 p = 0.28 None Regression of Q2 vs Q1 late firing Z scores for bilateral SCH-23390 Intercept 8O t −1.60 p = 0.13 None Slope 8O t 5.72 p = 3.17e-5 None Slope vs 1 8O t −8.61 p = 2.13e-7 None List of statistical comparisons and results.
In this issue
