Extended Data Fig. 13: Two-action learning analyses supplementing Fig. 5.
From: Dynamic behaviour restructuring mediates dopamine-dependent credit assignment
a, Linear relationship between initial T1→T2 interval and differential T1-T2 refinement, despite accounting for within-session changes and excluding data upon reaching rising phase of learning. Non-zero slope significance: p < 0.001. b, Action similarities between T1 and T2 do not account for variation in differential T1-T2 refinement. n = 14 ChR2-YFP mice (biological replicates). Black dashed line - linear fit. Grey dashed line - sigmoidal fit. c, Full plots of T1→T2 and T2→T1 median interval changes across learning stages. Repeated-measures 2-way ANOVA. time-specific difference (Slow Learners – F(2.184, 26.20) = 54.21, p = 5.3*10−10; Fast Learners – F(1.700, 20.40) = 92.12, p = 6.3*10−9). Post-hoc 2-tailed Tukey’s multiple comparisons tests. **** p < 0.0001. *** p < 0.001. ** p < 0.01. * p < 0.05. n.s. – not significant. See Supplementary Information for statistical/sample details.
