Figures and data
Go and No-go performance is similar across task variants.
a) Top: Schematic of the Go/No-go (short) behavioral task. Rats initiated trials by entering a nose-poke port (arrow, dotted line), either voluntarily (Self-initiated) or following nose-poke light illumination (Cue-initiated). After staying in the nose-poke port for 0.5s, an auditory stimulus (action cue) instructed rats to either initiate action (lever presses: ‘Go’, green) or remain in the nose-poke port (‘No-go’, red). In the Go/No-go/Free (long) task, an additional action cue was presented that did not require an action for rewards (‘Free’, blue). Following Free cue onset, rats could approach the reward magazine immediately. Bottom: Schematic of task timeline. For all trials, reward-delivery latency (relative to action-cue onset) was similar. Shaded gray area depicts approximate duration of action-cue onset for “short” and “long” task variants b-e) Success rates across Go/No-go task variants. f-i) Latency to orient toward the reward magazine following action-cue onset. j-m) Speed of movement toward the reward magazine after orientation. Significance: * p < 0.05, ** p < 0.01, *** p < 0.001, Wilcoxon matched-pairs signed rank or post hoc Dunn’s tests.
VMS dopamine encodes both action initiation and reward-magazine arrival irrespective of action contingency.
a) Histological verification summary of VMS electrode placements for all animals used in this study (n = 21). b) Representative color plot and corresponding dopamine trace in response to an unpredicted pellet. The circle indicates pellet delivery. Inset: Accompanying cyclic voltammogram confirms dopamine detection. c-e) Average dopamine concentration (nM; mean + SEM) for Go (green), No-go (red), and Free (blue) trials. Triangles denote the average latency-to-maximum dopamine relative to different behavioral epochs. Subject-wise comparisons of dopamine data were made for all alignments. d) Traces aligned to action-cue onset. Shaded gray area depicts approximate duration of action-cue onset for “short” and “long” task variants. Vertical stripes highlight differences in Go vs. No-go dopamine. Horizontal bars indicate timepoints with significant within-subjects differences between trial types. d) Traces aligned to nose-poke exit and e) magazine arrival were analyzed by comparing latency-to-max dopamine release. Box plots depict individual latencies to maximum dopamine release (± 2s) around magazine arrival for each animal. f) Dopamine release during the inter-trial interval (ITI) as rats approached the reward magazine from the opposite wall, aligned to magazine arrival. Boxplot statistical significance: * p < 0.05, ** p < 0.01, *** p < 0.001, Wilcoxon matched-pairs signed rank or post hoc Dunn’s tests.
Classified No-go behavior reveals VMS dopamine release following action-cue offset in ‘Biting’ trials.
a) Left: Agglomerative hierarchical clustering of behavior during the No-go action-cue period using Ward’s linkage and Euclidean distance. A three-cluster solution (dotted grey line) was selected resulting in a balanced distribution of trials across clusters and sessions. These clusters corresponded to three distinct behavioral strategies. ‘Biting’ - defined as biting of the nose-poke-port wall; ‘Digging’ - digging movements in the nose-poke port; ‘Calm’ - all other no-go trials. b) A supervised decision-tree classifier labeled trials to one of three behavioral groups: Digging, Biting, and Calm. Average dopamine concentration of each classified group is depicted (nM; mean + SEM). c) Maximum dopamine concentration (nM) following action-cue offset (+2s) was highest in Biting trials as compared to other No-go trials. ‘After action-cue offset: No-go (trials)’: Individual No-go trials classified by No-go behavior, and trial-wise statistical Kruskal-Wallis tests were performed for each task-variant. Solid lines depict the median. ‘After action-cue offset: No-go (rats)’, ‘After action-cue offset: Go (rats)’, and ‘Unpredicted pellets (rats)’: Rats classified based on their predominant No-go strategy, with no statistical tests performed. Solid lines depict the mean. Significance: * p < 0.05, ** p < 0.01, *** p < 0.001, post hoc Dunn’s tests.
Number of subjects per task variant with FSCV recordings and number of trials in each No-go behavioral classification.
Statistical analyses were performed between classified No-go trials due to limited sample sizes per group.
Rats did not continuously occupy the nose-poke port during self-initiated intertrial intervals (ITIs).
Percentage of time spent in the nose-poke port between trials of self-initiated Go/No-go and Go/No-go/Free task variants. Animals were permitted to enter the nose-poke port during the ITI but entry could not trigger action-cue presentation. The ITI for the self-initiated Go/No-go task was fixed at 5s, whereas for Go/No-go/Free task, it was varying between 15-25s.
Latency to initiate trials in Cue-initiated task versions.
The latency to initiate a trial following nose-poke light illumination was similar across trial types within task variants. Cue-initiated Go/No-go: W = 18, p = 0.47; Cue-initiated Go/No-go/Free: H(2) = 0.93, p =v 0.63. Horizontal dotted lines depict the median response latency to start trials after the cue turned on.
Latency to arrive at the reward magazine was not significantly different between Go and No-go trials.
After satisfying response requirements in Go and No-go trials, rats arrived at the reward magazine at similar time points. During Free trials, animals arrived at the magazine within 2s of hearing the action-cue tone. Action-cue onset is at 0s. Significance for post hoc Dunn’s tests: * p < 0.05, ** p < 0.01, *** p < 0.001.
Decision-tree classification of No-go trials.
Individual trials (n = 855) were categorised into one of three groups.
Classified No-go behaviors reveal no changes in dopamine during the action-cue epoch.
Column 1) Dopamine concentration during action-cue presentation did not differ significantly between no-go trial classifications (all p > 0.05; Kruskal-Wallis test statistics in Statistics table). Solid horizontal lines depict the median. Columns 2-4) Rats were grouped according to their predominant No-go strategy (see Methods; horizontal lines depict the mean). Column 2) Differences in average dopamine in No-go animals did not qualitatively differ between groups.
Biting trials were characterized by earlier nose-poke exits but inconsistent magazine approach speeds, ruling out a generalized vigor explanation for reward-approach dopamine changes.
Column 1) Latency to exit the nose-poke port from action-cue onset. Across Go/No-go/Free task variants, Biting trials consistently exhibited the shortest nose-poke hold times. Column 2) Speed to approach the reward magazine after nose-poke exit. While Kruskal-Wallis tests were statistically significant for all comparisons (p < 0.05; see supplementary statistics table), magazine approach speed showed no consistent relationship with Biting trials relative to other classifications. Together, the results rule out ‘vigor’ during magazine approach as the sole driver of reward-approach dopamine changes. Post hoc Dunn’s test significance: * p < 0.05, ** p < 0.01, *** p < 0.001.
VMS dopamine during Go trials tracks magazine arrival after the final lever-press and is independent of vigor.
For each task variant, trials were divided into quartiles based on last lever-press latencies (Q1 = fastest 25%; Q4 = slowest 25%). Column 1) When aligned to action-cue onset, the timing of maximum dopamine during reward approach differed between quartiles. Column 2) When realigned to magazine arrival, maximum dopamine was aligned to magazine arrival rather than lever press, as evidenced by consistent timing of maximum dopamine across quartiles. Column 3) Speed to approach the reward magazine after the last lever press. Kruskal-Wallis tests were statistically significant for Self-initiated Go/No-go and Cue-initiated Go/No-go/Free (p < 0.001; see supplementary statistics table). Comparisons of magazine approach speed within each task variant showed no consistent relationship between quartiles, ruling out post-action-cue ‘vigor’ as a main driver of reward-approach dopamine changes. Post hoc Dunn’s test significance: * p < 0.05, ** p < 0.01, *** p < 0.001.
Schematic of all VMS dopamine recording sites.
Coronal sections range from 0.48 to 2.28 mm anterior to Bregma. (Paxinos & Watson 1998).
Video analysis workflow.
Details for each step can be found in Methods Section 5: Video recordings and analysis - DeepLabCut pose estimation. Analysis details can be found under DeepLabCut behavior analysis.
Maximum dopamine concentration per trial as a function of distance to the magazine.
The x-axis represents neck distance from the magazine, progressing from furthest to closest (−2cm). Magazine arrival was defined as the moment neck distance reached <2 cm from the magazine-panel (dotted vertical line). The y-axis shows maximum dopamine concentration recorded within each trial. Individual dots represent the maximum dopamine concentration per trial. Bottom insert (histogram) shows the distribution of trial counts across distance bins. Note the accumulation of max dopamine values as animals arrive at the magazine (dotted line).
Maximum dopamine concentration around reward delivery (±3s).
The x-axis represents time (s), centered on reward delivery (time = 0 s). The y-axis shows maximum dopamine concentration recorded within each trial. Individual dots represent the maximum dopamine concentration per trial. The bottom insert (histogram) shows the distribution of trial counts across time bins. Note the accumulation of peak dopamine values prior to reward delivery in Free trials but not the other trial types.
Maximum dopamine concentration from the earliest predictor of reward to reward delivery.
The x-axis represents time (s), with time = 0 s depicting the earliest predictor of reward in each task variant. The y-axis shows maximum dopamine concentration recorded within each trial. Individual dots represent the maximum dopamine value per trial. The bottom insert (histogram) shows the distribution of trial counts across time bins. For Self-initiated task variants, time = 0 s corresponds to the onset of the action cue (speaker symbol). For Cue-initiated task variants, time = 0 s corresponds to illumination of the nose-poke light (lightbulb symbol). Note that for Cue-initiated Go/No-go/Free, 2 Go trials and 3 Free trials fell outside the x-axis range (0–10 s); for Self-initiated Go/No-go/Free, 7 Go trials and 10 Free trials fell outside this range.
