Dopaminergic mechanisms for motivation and reinforcement learning

点击翻译按钮获取中文摘要

Summary / Abstract This research program seeks to describe and understand striatal dopamine (DA) signals, particularly in the nucleus accumbens (NAc). NAc DA is a key regulator of both motivation and reward-related learning, and drugs of abuse share the common property of enhancing NAc DA. Yet we lack a clear account of how DA signals arise, how they are shaped by local NAc circuits, and how they modulate NAc output to influence behavior. Our Aims for the next funding period address three specific aspects of DA signals that are currently under active debate: First, why are NAc DA signals different to those in other parts of striatum? We found that NAc DA evolves more slowly compared to dorsal areas, and shows a distinct response profile to reward-predictive cues. We hypothesize that neural circuits involving NAc generate reward predictions and motivation over longer time horizons. Using retrograde optogenetic tagging, we will compare the firing patterns of individual DA cells that project to NAc subregions and other striatal areas, testing whether they define a temporal topography of reward expectation and feedback. Second, does NAc DA release convey signals beyond those encoded in the firing of afferent DA cells? We showed that changing the available reward for a behavioral task influences rats' willingness to work, and alters NAc DA release, without any apparent change in DA cell firing. We will investigate whether motivation- related changes in DA are instead sculpted by local cholinergic interneurons. To do this we will combine real- time DA and ACh measurements, selective pharmacological agents, and optogenetic tagging of cholinergic cells. Third, how do increases in DA release rapidly boost motivation? This is generally thought to involve enhanced firing of NAc output neurons that express the D1 DA receptor, relative to those expressing the D2 receptor. However, some recent results suggest that both populations increase activity, while computational models have cast doubt on whether DA can modulate cells quickly enough to explain observed behavioral effects. To resolve this we will record the spiking of identified NAc D1+ and D2+ output neurons, while also measuring and manipulating DA, at key behavioral moments. In each case we will take advantage of recent technical advances in optical chemical sensors, and of our ability to record the firing of individual identified neurons in freely-moving rats. We will use multiple carefully-controlled behavioral tasks, including an innovative new maze task in which rats display a trade-off between expected reward and required effort. Together these studies will provide vital new results for our understanding of NAc DA functions, along with rich, publicly-available data sets for the research community.