Modeling multi-area dynamics during motor control

阐明行为的神经基础是神经科学的基本目标。朝着这一目标取得的进展是 复杂的事实是，大多数行为都是由多个分布式和跨环境的 相互连接的大脑区域。项目4使用与实验数据紧密关联的模型来解决 顺序运动行为的背景。许多复杂的运动行为可以分解成以下序列 陈规陋习的成分或‘主题’，可以重新安排，以产生各种其他行为。要创建 作为现有基序的序列，运动系统必须在监测的同时产生所需的神经活动 运动进度，以便对不同主题之间的过渡进行适当的计时。这个建模项目是 旨在开发和测试代表运动皮质、运动丘脑和 基底节的输入和输出结构(即纹状体和GPI/SNR)，可以自主地产生 各种各样的基序，将它们灵活地串在一起fl成序列，并监控正在进行的活动，以确保 主题之间的转换在它们应该发生的时候发生。在这个模型中，大脑皮质和丘脑之间的环路产生 一个单一的皮质-丘脑网络，用于执行多个行为主题。关键的是，大脑皮质-丘脑 网络不是一个被抑制的实体，但可以被基底节的抑制性输出所改变(fifi)。这个主题就是 在任何给定时间产生的皮质-丘脑网络由运动丘脑中的哪一组神经元决定 不受当时GPI/SNR活性的抑制。不同的图案将通过改变图案来选择 改变GPI/SNR的活性，从而改变运动丘脑的抑制模式。因此，该组织的作用 模型中的GPI/SNR是为了保持当前基序，并驱动向序列中的下一个基序的过渡。 GPI/SNR的模型将被用来研究和提出它们在 主题，并在主题之间进行切换。纹状体将被模拟为皮质活动的监测器，其作用是阻止- 当一个主题已经结束而下一个主题可以开始的时候挖掘。当已经识别出适当的机会时，fi， 纹状体的短暂活动将触发系统通过其投射从一个基序切换到另一个基序 至GPI/SNR。这个建模项目与这个小组提案的总体目标紧密匹配， 对运动系统所产生的行为的定量理解。因为该模型将 神经群体在多个区域(运动皮质、运动丘脑、GPI/SNR和纹状体)的活动，它将 提供许多预测，这些预测将使用以下其他项目产生的实验数据进行测试 这个团体提案。这些测试的结果将被用于对模型进行再验证(fine)，此外还将用于预测 该模型的建立将指导新的实验方法。从实验和实验中得出约束的循环 来自模型的预测将继续下去，直到我们得出一个具有启发性和生物学上可信的电路级去... 解释了我们在实验中观察到的内容的脚本，并为fl如何扩展顺序提供了新的见解 生成运动序列。

Elucidating the neural basis of behavior is a fundamental goal of neuroscience. Progress towards this goal is complicated by the fact that most behaviors arise from interactions between a number of distributed and intercon- nected brain regions. Project 4 uses models that are tightly linked to experimental data to address this issue in the context of sequential motor behaviors. Many complex motor behaviors can be decomposed into sequences of stereotyped components or `motifs' that can be rearranged to produce a wide variety of other behaviors. To create a sequence from existing motifs, the motor system must generate the required neural activity while monitoring movement progress in order to time transitions between different motifs appropriately. This modeling project is aimed at developing and testing a model of interacting networks representing motor cortex, motor thalamus and input and output structures of the basal ganglia (i.e., striatum and GPi/SNr) that can autonomously generate a wide variety of motifs, string them together flexibly into sequences, and monitor ongoing activity to assure that transitions between motifs occur when they should. In this model, the loop between cortex and thalamus creates a single cortico-thalamic network for the execution of multiple behavioral motifs. Critically, this cortico-thalamic network is not a fixed entity but can be modified by the inhibitory output of the basal ganglia. The motif that the cortico-thalamic network produces at any given time is determined by which set of neurons in the motor thalamus is not being inhibited by GPi/SNr activity at that time. Different motifs will be selected by changing the pattern of activity in the GPi/SNr, thereby modifying the pattern of inhibition in the motor thalamus. Thus, the role of the GPi/SNr in the model is to maintain the current motif and to drive transitions to the next motif in a sequence. Models of the GPi/SNr will be used to study and propose mechanisms by which they sustain activity during a motif and switch it between motifs. Striatum will be modeled as a monitor of cortical activity with the role of deter- mining when one motif has ended and the next can begin. When an appropriate opportunity has been identified, transient activity in the striatum will trigger the system to switch from one motif to another through its projections to the GPi/SNr. This modeling project is tightly matched to the overall goal of this group proposal, a detailed, quantitative understanding of the production of behavior by the motor system. Because the model relates the activities of neural populations in multiple regions (motor cortex, motor thalamus, GPi/SNr and striatum), it will provide many predictions that will be tested using the experimental data produced by the other projects within this group proposal. The results of these tests will be used to refine the model and, in addition, predictions of the model will guide new experimental approaches. The cycle of deriving constraints from experiments and predictions from models will continue until we arrive at an illuminating and biologically plausible circuit-level de- scription that accounts for what we observe in experiments and lends new insight into how flexible sequential motor sequences are generated.