Midbrain dopamine neurons provide teaching signals for goal-directed navigation

摘要: Summary In naturalistic environments, animals navigate in order to harvest rewards. Successful goal-directed navigation requires learning accurately estimate location and select optimal state-dependent actions. Midbrain dopamine neurons are known be involved reward value learning1–13. They have also been linked learning, as they play causal roles place preference14,15 enhance spatial memory16–21. Dopamine therefore ideally placed provide teaching signals for navigation. To test this, we imaged neural activity mice learned a closed-loop virtual reality corridor lick report the location. Across phasic responses developed visual cues trial outcome that resembled prediction errors indicated animal’s of We observed development pre-reward ramping activity, slope which was modulated by both stage task engagement. The ramp correlated with accuracy licks next trial, suggesting sculpted accurate location-specific action during Our results indicate midbrain neurons, through their improving Highlights investigated Phasic reflected respect subjective A slow leading up over enhanced engagement Positive slopes were followed improved performance on subsequent trials, role