Agency assignment in visuomotor control: a dedicated mechanism to attribute visual feedback to motor commands

对大多数物种来说，快速准确的运动能力对生存至关重要。为了执行精确的目标运动，视觉运动系统必须对意想不到的变化做出适当的反应。要做到这一点，它必须同时监控目标和移动肢体的视觉信息。在自然环境中，由于需要忽略无关的视觉对象，这项任务变得复杂。尽管如此，视觉运动系统迅速对视觉反馈作出反应，这表明它自动标记了中央凹旁视觉中运动肢体的信息。我认为视觉反馈和运动指令之间的联系是通过一种专门的机制实现的：代理分配，将结果归因于其原因的过程。该项目的第一部分调查了机构分配和视觉注意力依赖于共享或单独过程的程度。这两种机制从复杂的视觉场景中提取相关信息，但在功能上存在根本差异。我将测试代理分配是否独立于视觉注意力的焦点。许多自然的动作都需要四肢同时运动。由于视觉信息往往是模糊的，很难确定哪个肢体引起了哪个视觉变化。因此，视觉运动系统必须连续地估计代理的来源。在项目的第二部分，我将在贝叶斯框架中对这个评估过程进行建模。使用双手触碰动作的行为实验将测试解释视觉和运动参数如何调节估计过程的模型预测。项目的第三部分探索使用脑磁图代理分配的神经关联。我将测试代理分配取决于视觉和感觉运动区域神经活动的一致性这一假设。因此，这个项目建立了一个关于大脑如何在自然运动中实现有效反馈控制的新概念。

For most species, the ability to produce fast and accurate movements is critical for survival. To execute precise goal-directed movements, the visuomotor system must react appropriately to unexpected changes. To do this, it must simultaneously monitor visual information about the target and about the moving limb(s). In natural environments, this task is complicated by the need to ignore irrelevant visual objects. Nonetheless, the visuomotor system promptly reacts to visual feedback, suggesting that it automatically tags information about the moving limb(s) in parafoveal vision. I propose that the link between visual feedback and motor commands is achieved by a dedicated mechanism: agency assignment, the process of attributing outcomes to their causes. The first part of the project investigates the degree to which agency assignment and visual attention rely on shared or separate processes. Both mechanisms extract relevant information from complex visual scenes but differ fundamentally in their functions. I will test whether agency assignment operates independently of the focus of visual attention. Many natural actions involve simultaneously moving limbs. Because visual information is often ambiguous, it is difficult to determine which limb caused which visual change. Thus, the visuomotor system must continuously estimate the source of agency. In the second part of the project, I will model this estimation process in a Bayesian framework. Behavioural experiments using bimanual reaching movements will test the model predictions that explain how visual and motor parameters modulate the estimation process. The third part of the project explores the neural correlates of agency assignment using magneto encephalography. I will test the hypothesis that agency assignment depends on the coherence of neural activity in visual and sensory-motor regions. Thus, this project establishes a novel conception of how the brain achieves efficient feedback control during natural movements.