Forebrain dopamine value signals arise independently from midbrain dopamine cell firing

The mesolimbic dopamine projection from the ventral tegmental area (VTA) to nucleus accumbens (NAc) is a key pathway for reward-driven learning, and for the motivation to work for more rewards. VTA dopamine cell firing can encode reward prediction errors (RPEs 1,2 ), vital learning signals in computational theories of adaptive behavior. However, NAc dopamine release more closely resembles reward expectation (value), a motivational signal that invigorates approach behaviors 3-7 . This discrepancy might be due to distinct behavioral contexts: VTA dopamine cells have been recorded under head-fixed conditions, while NAc dopamine release has been measured in actively-moving subjects. Alternatively the mismatch may reflect changes in the tonic firing of dopamine cells 8 , or a fundamental dissociation between firing and release. Here we directly compare dopamine cell firing and release in the same adaptive decision-making task. We show that dopamine release covaries with reward expectation in two specific forebrain hotspots, NAc core and ventral prelimbic cortex. Yet the firing rates of optogenetically-identified VTA dopamine cells did not correlate with reward expectation, but instead showed transient, error-like responses to unexpected cues. We conclude that critical motivation-related dopamine dynamics do not arise from VTA dopamine cell firing, and may instead reflect local influences over forebrain dopamine varicosities.