Extended Data Figure 2-1
A, Sample image from DeepLabCut as the mouse is running on the treadmill. Each paw, the head, the base of the tail, and the tail are labeled by colored dots. B, Coordination score from treadmill running for DMS-injected and DLS-injected animals. Mice performed significantly better within 2–4 d from day 1 of training, indicating that surgery, viral expression, and optic fiber implant did not affect running abilities and motor learning (one-way ANOVA **p < 0.01 for DMS mice and *p < 0.05 for DLS mice). C, D, Average calcium event peak amplitudes for D1-SPNs in the DMS (C) and DLS (D) over the 12 d of training. Peak epoch amplitudes are compared with the baseline home cage recording. There is no significant change in home cage baseline (one-way ANOVA p = 0.13 for DMS and p = 0.73 for DLS) and time-off (one-way ANOVA p = 0.21 for DMS and p = 0.50 for DLS). Average peak amplitude for DMS D1-SPNs significantly decreased during running time (one-way ANOVA p < 0.0001 for on-time alone, * represents post hoc significant values, and RM two-way ANOVA p < 0.01 when comparing on-time to off-time, # represents post hoc significant values; detailed statistics in the main text) although peak amplitudes average levels remained unchanged for DLS D1-SPNs (one-way ANOVA p = 0.64). E, F, Average calcium event rate for D1-SPNs in the DMS (E) and DLS (F). Measurements for epochs of running time (treadmill on), time in which the treadmill is off, and event counts in home cage baseline recording were overlapped. There were no significant changes in the average event rate over the 12 d of training for DMS (one-way ANOVA p = 0.08 for home cage baseline recording, p = 0.07 for running time, and p = 0.31) or DLS recording (one-way ANOVA p = 0.67 for home cage baseline recording, p = 0.59 for running time, and p = 0.93). Download Figure 2-1, TIF file.