Fig. 3: Low dimensional population activity alignment changes between early and late learning. | Nature Communications

Fig. 3: Low dimensional population activity alignment changes between early and late learning.

From: Dissociable roles of distinct thalamic circuits in learning reaches to spatial targets

Fig. 3: Low dimensional population activity alignment changes between early and late learning.

a PfFL trial-averaged fluorescence aligned to movement start for matched cells on two early days of block 1 (n = 113 cells, 5 mice), two late days of block 1 (n = 114 cells, 5 mice), two days around block switch (n = 116 cells, 5 mice), two early days of block 2 (n = 67 cells, 4 mice), and two late days of block 2 (n = 49 cells, 4 mice). Wilcoxon matched-pairs signed rank test, two-sided: early 1–2 (block 1) *p < 0.0001, late 1–2 (block 1) p > 0.05, late 2 (block 1) – early 1 (block 2) *p < 0.0001, early 1–2 (block 2). *p < 0.0001, late 1–2 (block 2) p > 0.05. Data shown as mean ± SEM. b Same as in (a) for VALFL. Two early days of block 1 (n = 46 cells, 4 mice), two late days of block 1 (n = 72 cells, 5 mice), two days around block switch (n = 86, 5 mice), two early days of block 2 (n = 97 cells, 5 mice), and two late days of block 2 (n = 103 cells, 5 mice). Wilcoxon matched-pairs signed rank test, two-sided: early 1–2 (block 1) *p = 0.0008, late 1–2 (block 1) p > 0.05, late 2 (block 1) – early 1 (block 2) p > 0.05, early 1–2 (block 2). p > 0.05, late 1–2 (block 2) *p = 0.01. c PfFL population distribution of positively (white), negatively (gray), and non-modulated (dark gray) cells on all matched day pairs. Fisher’s exact test (performed on cell counts): E1-E2 (block 1) *p = 0.013, E1-E2 (block 2) *p = 0.045. All other matched day pairs’ Fisher’s exact tests: p > 0.05. d Same as in (c) for VALFL. Fisher’s exact test for all matched day pairs: p > 0.05. e Top 3 PCs of neural activity of PfFL matched cell populations around movement start for late day of block 1 and early day for block 2. f Same as (d), for VALFL matched neuronal populations. g For PfFL, average alignment of subspaces occupied within (open circle with dot) and across (filled circle) days with matched cells from two early days (E) in each block (black, average over block 1 and 2), two late days (L) in each block (magenta, average over block 1 and 2), and the block switch day from late block 1 to early block 2 (L/E, blue). For PfFL (n = 5 mice), two-tailed paired t test within vs across; E, *p = 0.026; L, p > 0.05; L, E, *p = 0.0015. center and error bars indicate mean ± SEM. h Same as (g), for VALFL (n = 5 mice) matched cell populations. Two-tailed paired t test within vs across; E, *p = 0.0008; L, *p = 0.0037; L/E, *p = 0.0073. Error bars shown as mean ± SEM. See also Supplementary Fig. 6–8. Source data are provided as a Source Data file.

Back to article page