Neuromodulation of flexible foraging

我是该留下还是该走？是一种我们都会反复面对的决定动物如何解决这种困境一直是众多学科的兴趣所在，包括心理学、经济学、人工智能和行为生态学。引人注目的是，通过研究动物觅食行为，我们找到了一个行之有效的解决方案：只有当一种动物的回报率高于环境中其他地方的平均回报率时，它才能继续坚持当前的策略。这一框架被称为边际价值定理（MVT），已被证明在解释各种环境和动物物种的去留决定方面非常有效，从无脊椎动物的觅食到人类的信息搜索。尽管如此，大脑是如何有效地跟踪关键变量以实现这一点的，目前还不清楚。在神经科学中，人们普遍认为，多巴胺和血清素等大脑化学物质在协调适应性行为中起着至关重要的作用。例如，破坏这些化学物质，无论是实验性地使用药物制剂，通过健康衰老的自然过程，还是精神或神经疾病的结果，都可以改变动物适应环境变化的速度。重要的是，我们对这些化学物质作用的理解主要来自简单的决策任务，即动物在选项之间重复选择，并根据试错经验更新其行为。因此，我们知识的一个主要空白是这些化学物质在指导更自然的动态觅食决策中所起的作用。有趣的是，我们团队最近的一些工作表明，多巴胺可以同时发出两个不同变量的信号，这两个变量可能对这种觅食决策很重要：（1）多巴胺的瞬时波动（持续<1秒），跟踪预期和获得的奖励之间的差异，被认为是学习奖励值的基本教学信号，以及（2）多巴胺的持续变化（持续数秒或更长时间），跟踪环境中的平均潜在增益。此外，最近的方法学进步已经彻底改变了我们监测大脑化学物质每时每刻变化的能力，以及以亚秒级精度操纵它们的能力，这两项发展为我们第一次解决多巴胺和血清素在执行有效的动态留或走决策中扮演的角色奠定了基础。首先，我们将使用一种称为纤维光度法的光学技术来测量执行觅食任务的小鼠大脑中的多巴胺水平。通过操纵（i）它们在觅食地获得奖励的频率，以及（ii）它们到达另一个觅食地需要多长时间，我们可以确定多巴胺的每时每刻的变化如何与它们留下或离开的决定相关，以及这些变化如何与模型的预测相一致。为了做到这一点，我们将使用一种叫做光遗传学的技术，在老鼠执行觅食任务时选择性地打开或关闭表达多巴胺的神经元;第三，我们将研究另一种大脑化学物质血清素是否通过调节多巴胺的释放来影响留下或离开觅食的决定。为此，我们将使用新的技术组合来操纵血清素，同时在小鼠执行觅食任务时监测多巴胺，使用与上述相同的技术。我们将验证我们的假设，即通过调节多巴胺的释放来促进对奖励的反应，从而提高血清素的水平，这将为多巴胺如何与血清素相互作用，从而在自然环境中实现有效的灵活性提供重要的见解。

'Should I stay or should I go?' is a type of decision that we all repeatedly face. How animals solve this type of dilemma has been of interest across a wide range of disciplines, including psychology, economics, artificial intelligence and behavioural ecology. Strikingly, a well-established efficient solution to the problem emerged from studying animal foraging behaviour: an animal should only continue to persist with its current strategy if it yields greater returns than the average of the opportunities available elsewhere in the environment. This framework, called the Marginal Value Theorem (MVT), has proven remarkably effective in explaining stay-or-leave decisions across a wide variety of settings and animal species, from foraging for food in invertebrates to searching for information in humans. Nonetheless, it is not clear how the brain keeps track of the key variables to enable this to happen efficiently.In neuroscience, it is generally accepted that brain chemicals such as dopamine and serotonin play essential roles in coordinating adaptive behaviour. For example, disrupting these chemicals, either experimentally using pharmacological agents, through natural processes in healthy ageing, or as a consequence of psychiatric or neurological disease, can change how quickly animals adapt to changes in their environment. Importantly, our understanding of these chemicals' roles comes mainly from simple decision making tasks where animals make repeated choices between options and update their behaviour based on trial-and-error experience. Therefore, a major gap in our knowledge is the roles these chemicals play in guiding more naturalistic dynamic foraging decisions.Intriguingly, some of our team's recent work has demonstrated that dopamine can simultaneously signal two different variables that could be important for such foraging decisions: (1) transient fluctuations in dopamine (lasting <1 second) that track the difference between the expected and obtained reward, believed to be an essential teaching signal for learning reward values, and (2) sustained changes in dopamine (lasting multiple seconds and longer) that track the average potential gains in the environment. Moreover, recent methodological advances have revolutionised our ability to monitor moment-by-moment changes in brain chemicals as well as to manipulate them with sub-second precision.These two developments set the stage for us to address for the first time what role dopamine, in concert with serotonin, plays in implementing efficient dynamic stay-or-leave decision making. First, we will measure dopamine levels in the brains of mice performing a foraging task using an optical technique called fibre photometry. By manipulating (i) how often they get rewards in the patch they are foraging in and (ii) how long it will take them to reach an alternative foraging site, we can determine how moment-by-moment changes in dopamine correlate with their decisions to stay or leave and how these align with predictions from models.Second, we will test whether these changes in dopamine play a causal role in animals' choices to stay or leave. To do this, we will use a technique called optogenetics to selectively turn on or off neurons that express dopamine as the mice perform the foraging task.Third, we will investigate whether another brain chemical, serotonin, influences stay-or-leave foraging decisions by modulating dopamine release. To do this, we will use new combinations of techniques to manipulate serotonin while simultaneously monitoring dopamine as mice perform the foraging task, using the same techniques as described above. We will test our hypothesis that boosting serotonin promotes persistence by modulating dopamine release in response to rewards.Together, this will provide important insights into how dopamine interacts with serotonin to enable efficient flexibility in naturalistic environments.