|
Fig. 4 Notations in the figures are all the same as in Fig.3b. a The activity of two neural ensembles (cyan and magenta lines) in the adaptation stage (left panel) and the stages immediately before (middle panel) and after (right panel) behavioral learning was established. b Enlarged view of the activity of the two ensembles in the boxed area in (a). The vertical gray line indicates the time point when the fish reached the goal. c Comparison of the cyan ensemble?s peak activity in the GO trials in different learning stages. Columns and error bars: mean ± SEM. Each circle indicates the value in each GO trial. The number in parentheses is the number of trials used in the statistics. Adaptation vs immediately before behavioral learning, ***P = 2.60 × 10?16; adaptation vs after behavioral learning, ***P = 6.43 × 10?13, F(3, 55) = 82.85. One-way ANOVA, Bonferroni?s multiple comparison test. d Comparison of the magenta ensemble?s peak activity when fish perceived red color in GO and NOGO trials in different learning stages. Columns and error bars: mean ± SEM. Each circle indicates the value in each trial. The numbers in parentheses are the number of trials used in the statistics. Red GO before behavioral learning vs Red GO after behavioral learning, **P = 7.91 × 10?3; red NOGO before behavioral learning vs red NOGO after behavioral learning, **P = 6.48 × 10?3, F(7, 80) = 9.25. One-way ANOVA, Bonferroni?s multiple comparison test. e?h Results of the same analysis as (a?d) above for another fish. g The numbers in parentheses are the number of trials used in the statistics. Adaptation vs immediately before behavioral learning, **P = 1.94 × 10?3; adaptation vs after behavioral learning, ***P = 2.44 × 10?5, F(3, 56) = 13.02 One-way ANOVA, Bonferroni?s multiple comparison test. h The numbers in parentheses are the number of trials used in the statistics. Red GO adaptation vs red GO intermediate stage, P = 0.295; red NOGO adaptation vs red NOGO intermediate stage, ***P = 9.085 × 10?11, F(7, 88) = 31.67. One-way ANOVA, Bonferroni?s multiple comparison test. Red GO adaptation vs red GO intermediate stage, **P = 1.42 × 10?3. Two-tailed unpaired t-test. i Relationship between the correlation coefficient and distance among the 10 most-contributing neurons in the color rule-encoding ensembles (left, blue is dangerous rule; right, red is safe rule). The data were averaged from the fish with these ensembles. Black line denotes the averaged correlation from all fish. Red line denotes the average of averaged 10 shuffled data from the fish (see ?Methods?).