Investigating a role for dopamine in organizing behavioral sequencing

摘要 大脑评估环境，做出决定，并将决定转化为行为。行为，还是总数 生物体所表现出的向外运动，是通过将小的、定型的碎片缝合在一起而构成的。 动作，或动作的“音节”。动物先天行为的复杂性部分源于这样一个事实， 动物能发出的几十个或几百个音节可以灵活地连接成几乎无限的音节， 音节序列的变化。有些序列是确定性的--音节B总是跟在音节A之后 在时间上-而其他人更概率地构建。例如，逃跑行为通常更多 确定，而觅食是以更概率的方式组成的。在哺乳动物的大脑中， 纹状体（DLS）负责将音节连接成行为序列，但尚不清楚纹状体是如何将音节连接成行为序列的。 DLS将动作串成可变的或原型化的序列。一种假设是， DLS中的神经递质多巴胺能够携带信息，并可能调节 DLS选择构建行为序列。然而，测试这个假设需要开发一个 连续的行为如何被分割成离散的音节运动的定量描述。最近的工作 在我们的实验室已经开发出一种算法称为运动测序，或简称为'MoSeq'，它客观， 自动将鼠标行为分割成其组成音节。因为MoSeq能给我们精确的时间尺度 关于每个音节的身份以及一个音节与该音节之间的时间边界的信息。 接下来，我们现在可以探索DLS中的多巴胺活性如何与正在进行的小鼠的序列结构相关。 行为在这里，我建议将MoSeq的分割管道与神经成像和操作相结合 研究纹状体多巴胺和行为序列变异性之间关系的方法。基于 根据初步数据，我假设高水平的多巴胺释放既预示着 可变序列，并且在自发行为期间因果地影响序列可变性。在Aim中 1，我将同时在DLS和MoSeq行为记录中记录多巴胺释放;我将 研究如何音节相关的多巴胺与行为序列的变异性，其中一个 音节是嵌入的。在目标2中，我将利用MoSeq平台的实时变体，其中音节可以 当一个动物在表演它时，我会刺激DLS投射多巴胺能轴突， 特定目标音节的表达，并检查目标之后音节的变化。 这些实验将揭示纹状体多巴胺和行为序列之间的快速时间尺度关系 并将促进我们对动物如何能够灵活地构建音节序列的理解。 每时每刻的基础上。

Abstract The brain assesses the environment, makes decisions, and turns decisions into behavior. Behavior, or the total outward movement displayed by an organism, is constructed by stitching together small, stereotyped bits of movement, or ‘syllables’ of action. The complexity of innate animal behavior is in part derived from the fact that the dozens or hundreds of syllables that an animal can perform can be flexibly concatenated into a nearly infinite variety of syllable sequences. Some sequences are deterministic — with syllable B always following syllable A in time — while others are constructed more probabilistically. Escape behaviors, for example, are typically more determined, while foraging is composed in a more probabilistic manner. In the mammalian brain, the dorsolateral striatum (DLS) is responsible for concatenating syllables into sequences of behavior, but it is unclear is how the DLS strings actions together into either variable or stereotyped sequences. One hypothesis is that the release of the neurotransmitter dopamine in DLS is capable of carrying information about – and perhaps regulates – how the DLS chooses to construct behavioral sequences. However, testing this hypothesis requires developing a quantitative description of how continuous behavior is segmented into discrete syllable movements. Recent work in our lab has developed an algorithm called Motion Sequencing, or ‘MoSeq’ for short, which objectively and automatically segments mouse behavior into its component syllables. Since MoSeq affords us fine-timescale information regarding the identity of each syllable and the temporal boundaries between one syllable and the next, we can now explore how dopamine activity in the DLS relates to the sequence structure of ongoing mouse behavior. Here, I propose to combine the MoSeq’s segmentation pipeline with neural imaging and manipulation approaches to investigate the relationship between striatal dopamine and behavioral sequence variability. Based upon preliminary data, I hypothesize that high levels of dopamine release both predicts the expression of variable sequences, and causally influences sequence variability during spontaneous behavior. In Aim 1, I will simultaneously perform recordings of dopamine release in DLS and MoSeq behavioral recordings; I will investigate how syllable-associated dopamine relates to the variability of the behavioral sequence in which a syllable is embedded. In Aim 2, I will leverage a real-time variant of the MoSeq platform, in which a syllable can be detected as an animal is performing it. I will stimulate DLS-projecting dopaminergic axons during the expression of specific target syllables and examine changes in the variability of syllables that follow the target. These experiments will reveal fast-timescale relationships between striatal dopamine and behavioral sequences and will advance our understanding of how animals are able to flexibly construct sequences of syllables on a moment-to-moment basis.