Fig. 6: Population activity dynamics of the mThLCN→M2 axons that represent the difference between post-success and post-error trials in H sessions.
a, b Trajectory of PC1–PC3 of mThLCN→M2 axonal activity in post-success (blue) and post-error (orange) trials in H sessions. Open circles, filled stars, and filled squares show cue onset, cue+0.2s, and average pull-onset time points, respectively. Axons in sessions with at least one of each of the six types of trials were used (316 axons from 14 sessions). b Trials were divided into three groups according to pull latency (left, 700–900 ms; middle, 900–1100 ms; right, 1100–1300 ms). c Euclidean distance in the trajectory of the top five PCs between post-success and post-error trials at cue–0.4s, cue+0.2s, pull–0.3s, and pull-onset time points. Trials were divided into three groups according to pull latency (green, 700–900 ms; blue, 900–1100 ms; purple, 1100–1300 ms). Asterisks indicate values above the 99th percentile of chance levels (dashed lines), which were estimated by shuffling the trial assignments and performing the same calculation. d Prediction accuracy of the population activity for classifying the trials into post-success and post-error trials. n = 14 sessions. Black bars, the mean ± SEM. **P = 0.00122, compared with the value at the cue–0.4s time point, Wilcoxon signed-rank test with Bonferroni correction (α = 0.0167), two-sided. +++P = 1.22 × 10–4 for cue-0.4s, cue+0.2s, pull−0.3s, and pull-onset time points, compared with 0.5, Wilcoxon signed-rank test with Bonferroni correction (α = 0.0125), two-sided. e Matrices of Pearson correlation coefficients for the population vectors (316 axons from 14 H sessions) that represented the one-back trial outcome between each time. The vectors were aligned to the cue onset (left) and pull onset (right). f Proportions of the axons that were modulated by the one-back trial outcome at each time point and at both of the two time points. Asterisks indicate statistically significant proportions (more than the 99th percentile of the shuffled data). g Schematic of transitions of the one-back trial outcome-representing axons (red arrows). Black arrows indicate the axons that did not represent it. Thick gray arrows indicate the axons that represented the one-back trial outcome across different time points. Source data are provided as a Source Data file.
