Fig. 8: BF responses to trial outcome were negatively correlated with BF activities in earlier epochs.

a Average responses of BF bursting neurons (mean ± s.e.m.) to the reward in light lick trials, relative to their respective baseline firing rates, plotted separately for the five sessions relative to the D₂ session (left). BF responses to the reward (yellow shaded interval, left panel) were negatively correlated (Pearson correlation) with BF responses to light onset (Fig. 7c) in individual sessions (right). Each circle indicates one session and different colors correspond to different animals. b, Negative correlation of BF activities between evaluation and outcome responses in single trials. b1 Single trial BF activities in an example D₁ session, aligned at the trial outcome. Single trial BF responses were plotted separately for rewarded licks (light licks) and non-rewarded licks (catch licks and no-fixation licks). Yellow shaded intervals indicate time windows for calculating evaluation and outcome responses. Trials were sorted by the amplitude of evaluation responses (top). Average BF activities from the four quartiles of trials were plotted separately (bottom). b2 Negative correlation (Pearson correlation) between single trial BF evaluation responses and outcome responses in this session. Each dot represents one trial. Catch licks and no-fixation licks were pooled together to calculate the correlation in non-rewarded licks. c Histogram of correlation coefficients (Pearson correlation) between evaluation and outcome responses from individual sessions. Results for rewarded lick trials (light licks) were calculated from pre-D₂ sessions when BF responses to the light stimulus had not developed (N = 7), and for non-rewarded lick trials from sessions with at least 50 trials of catch and no-fixation licks combined (N = 25). Most sessions showed significant negative correlations.