Fig. 1: Multiphasic dynamics of DA and Ach in the VLS during reward-based decision-making.

a, Proposed DA and Ach interactions. CINs release Ach, which evokes the release of DA through nicotinic acetylcholine receptors (nAchRs) on DAN terminals. Conversely, DA inhibits CINs through D2Rs. VLS, ventrolateral striatum. b, 2ABT parameters. An LED (yellow) signals trial initiation. A single reward is probabilistically delivered when the mouse makes the correct choice (P(R|left) is the probability of a reward delivered at the left port and P(R|right) is the probability of a reward delivered at the right port), and the trial is terminated with side-port exit. Fibre photometry is simultaneously performed. c, Ipsilateral DA and Ach dynamics and licks (yellow) recorded from an example mouse during a 2ABT session. Each row depicts the z-scored sensor signal of a trial. d, DA and Ach signals during different reward outcomes. The averaged z-scored signal ± s.e.m. is shown. Data are aligned to side-port entry (SE) (DA: n = 13 mice; Ach: n = 14 mice). e, Statistical analysis of d. Rewarded versus unrewarded trials are compared (left and right dot in connected pair, respectively). Open circles denote a significant difference for each comparison (two-sided t-test (P < 0.05)). Mean DA and ∆Ach ± s.d. are shown. Percentages represent LDA classification accuracy. f, DA and Ach release during rewarded 2ABT trials in which the LEDs that signal centre-port and side-port entry are present or omitted. Data are shown as in d (n = 4 mice). g, DA and Ach release for trials in which mice choose the same port in both the previous and the current trials, which are segregated by the reward outcome of the previous trial. Data are shown as in d (DA: n = 13 mice; Ach: n = 14 mice). h, DA and Ach release for trials in which mice switch ports. In g,h, the text denotes previous outcome followed by current outcome (for example, Win, win). Data are shown as in d (DA: n = 13 mice; Ach: n = 14 mice).

Source Data