Fig. 4: Rapid emergence and subsequent refinement of reward-seeking behaviors and BF activity.
a The pattern of rightward licking behaviors aligned at the transition point in the D1 session of each animal (N = 6). One animal (#7) with accelerated learning dynamics, in which its D1 and D2 occurred in the same session, was excluded from this analysis. The behavioral response patterns in three trial types (light trials, catch trials and no-fixation licks) were either pooled together within each animal (a1), or plotted separately (mean ± s.e.m.) (a2). While all three types of rightward licking emerged immediately after the transition, no-fixation licks subsequently decreased after 60 trials (a3) (Repeated measures ANOVA for group comparisons and post-hoc two-sided paired t-test between two trial types, N = 6). b Corresponding changes in BF activity aligned at the transition point in the D1 session of each animal (N = 6). BF evaluation responses and outcome responses in the three trial types were pooled within each animal (b1), as in the example in Fig. 3e2. BF evaluation responses were further plotted separately for each trial type, relative to their respective baseline firing rates (mean ± s.e.m.) (b2). BF evaluation responses increased similarly in the three types of rightward licking immediately after the transition, and subsequently decreased in no-fixation licks after 60 trials (b3) (Repeated measures ANOVA for group comparisons and post-hoc two-sided paired t-test between two trial types, N = 6). Thin lines in b1 indicate the trend (10-trial moving median) of BF activities from individual animals.
