Fig. 3: Abrupt transition in reward-seeking behaviors corresponded to increased neuronal activity in the BF during initial learning.

a, Locations of electrode bundles targeting bilateral BF (N = 7) in coronal sections of the rat brain (coordinates relative to Bregma). Different colors correspond to different animals. Used with permission of Elsevier, from The rat brain in stereotaxic coordinates, Paxinos and Watson³⁴; permission conveyed through Copyright Clearance Center, Inc. b, Response of individual BF bursting neurons (n = 1013) to the S^left sound during new learning sessions (N = 45 sessions; separated by thin red lines) in each animal (N = 7 rats; separated by cyan dotted lines). BF bursting neurons showed robust and consistent phasic responses to the S^left sound throughout the learning process. c, Average BF bursting neuron responses to S^left sound onset and the associated reward delivery. BF activities in catch trials were plotted for comparison. Responses from individual animals (thin lines) were similar (N = 7). Symbols were adapted from Avila and Lin²². d The activity of BF bursting neurons remained stable between the first (D₀) and last (D_last) recording session. Average activities in the yellow shaded intervals were similar between these two sessions (inset) (two-sided paired t-test, N = 7). Thin lines indicate BF activity in individual animals. Symbols were adapted from Avila and Lin²². e Behavioral and BF neuronal dynamics in the D₁ session from a representative animal (rat #4). e1 The emergence of three types of rightward licks after the transition point (top), and their combined rightward licking probability across trial types (bottom). The transition point (T) marked an abrupt transition in the pattern of reward-seeking behavior that went from no licking to 100% licking. e2 Top, population activities of BF bursting neurons (color-coded) in the same trials (X-axis) as shown in e1. Y-axis indicates time in each trial, with time zero aligned at the trial outcome. No lick trials before the transition were aligned instead at the time of fixation port exit (fix-out, white circle) such that the median timing of fix-outs in lick and no lick trials were equivalent. The blue and red lines to the right of the panel indicate the time windows for calculating evaluation and outcome responses, respectively. Bottom, BF evaluation responses (blue) and outcome (red) responses across trials. Outcome responses were plotted separately for rewarded (solid red) and non-rewarded (dashed red) licks. Circles indicate BF activities in single trials and lines indicate their respective trends (moving medians). e3 Examples of single trial BF activities from the three types of rightward licks taken around the transition point. Each panel showed the spike rasters of BF bursting neurons in this session (n = 15) (top), along with the population activity trace and relevant behavioral events (bottom). Shaded intervals indicate the time windows corresponding to evaluation responses (blue) and outcome responses (red) shown in e2. Notice that BF bursting neurons in the same session showed highly similar activity patterns, and that the BF evaluation response rapidly emerged in all three types of rightward licks after the transition point.