A. Cumulative response curves for learners (black, n=11) and non-learners (red, n=10). B. Peak values of cue-evoked DA release are similar for learners and non-learners preceding reversal presentation. After reversal, cue-evoke DA continues to decrease for non-learners, but stabilizes for learners. Learners – black, non-learners – red. Blue line shows average peak value during first session of discrimination learning for comparison.
C. Learners and non-learners show similar motivation to lever press as indicated by similar response latencies before reversal. Learners – black bars, non-learners – open bars. D. Learners and non-learners show similar DA release to unexpected reward delivery but DA release to cues differs. Left – response to reward delivery averaged across the first two correct responses after reversal. Right – cue-evoked DA is updated after positive feedback in learners, but not in non-learners. Bar graph shows mean difference scores (cue-evoked DA on correct+1 trials – cue evoked DA on correct trials) for the first two correct responses after reversal. E. Changes in DA during first two correct responses after reversal. Heat plots show average DA values per trial following reward delivery (left panel) and cue presentation (center and right panels) for the first two correct trials (correct) and for the trial immediately following correct trials (correct +1). Upper panels show results for learners, lower panels show results for non-learners. Left panel – response to reward delivery, middle panel – response to cue presentation on correct trials, right – response to cue presentation on correct+1 trials. F. Positive correlation between percentage correct and cue-evoked DA release after reversal across all animals. Cue-evoked DA after reversal was normalized to cue-evoked DA release on last 10 trials before reversal to control for individual differences. Red dots indicate non-learners, black dots indicate learners. G. Changes in DA during first two incorrect responses after reversal. Left – response to lever press averaged for first two incorrect responses after reversal. Right – For cue-evoked responses difference scores between the first incorrect trials (error trials; negative feedback received) and the first trials on which this negative feedback could be used (error+1 trials) were not different between learners and non-learners.