A Look into the Future: Anticipatory saccades shed light on human action control

在日常生活中，人类的行为目标直接，没有明显的努力。从无数可能的动作中，我们准确地选择将产生所需效果的动作。内生行为控制的表意运动理论解决了这样一个问题，即人类如何能够通过提出行为与其在环境中的影响之间的双向关联来选择适当的行为以达到预期的效果，例如，在按下灯开关和打开灯之间。因此，预测效果允许选择将产生该效果的相关动作。在过去的几十年里，已经有大量的研究支持这样的观点，即行动效果预期通过影响行动选择来实现目标导向的行动控制。有趣的是，早期的意念运动理论家已经提出了人类行为控制的第二种核心机制：效果监控。然而，人类需要比较预期效果和实际效果来评估目标实现的这一主张，在目前关于人类行为控制的研究中得到的关注相对较少。最近，我开发了一种眼球跟踪范式，它允许同时评估预期行动选择和主动效果监控的过程。参与者按下左侧/右侧键，以便在短暂延迟后在左侧/右侧产生可预测的视觉动作效果。在这一范式中，可以通过产生空间相容和空间不相容效应的动作之间的性能差异来评估预期动作选择的过程，而通过对S效应出现之前发生的动作的可预测未来位置的预期扫视来评估效果监控。最有趣的是，这些预期眼跳的存在表明了一个反映预期效果的主动效果监测过程，并支持稍后对预期效果和实际效果的比较。在这个项目的范围内，我的目标是同时和系统地研究目标导向行动控制的两种机制--预期行动选择和主动效果监测--以及它们之间的相互作用，以获得一个统一的行动控制框架。我将首先研究在不同动作控制模式下更密切地反映在预期眼跳中的主动效果监控。随后，我将系统地研究未来行动效应的哪些特征被主动监测，以确定主动效应监测的特异性。最后，我将在一个提出的人类内生行为控制的统一模型中研究预期行为选择和主动效果监控过程之间的相互作用。

In everyday life, humans act goal-directedly without apparent effort. From an infinite number of possible actions, we select exactly the action that will produce a desired effect. Ideomotor theory of endogenous action control addresses the question how humans are able to select appropriate actions to achieve desired effects by suggesting bi-directional associations between actions and their effects in the environment, for instance, between pressing the light switch and the light turning on. Thus, anticipating an effect allows for selecting the associated action that will produce it. The last decades have seen a vast body of research supporting the idea that action effect anticipations enable goal-directed action control by influencing action selection. Interestingly, early ideomotor theorists already proposed a second mechanism central to human action control: Effect monitoring. This claim that humans need to compare expected and actual effects to evaluate goal achievement, however, has received relatively little attention in present research on human action control. Recently, I have developed an eye tracking paradigm that allows for simultaneously assessing processes of anticipatory action selection and proactive effect monitoring. Participants press left/right keys in order to produce predictable visual action effects on the left/right side that occur after a short delay. In this paradigm, processes of anticipatory action selection can be assessed via performance difference between actions that will produce spatially-compatible and spatially-incompatible effects, whereas effect monitoring is assessed via anticipatory saccades towards the predictable future location of an action´s effect that occur before this effect has appeared. Most interestingly, the presence of these anticipatory saccades suggests a proactive effect monitoring process that reflects the expected effects and supports a later comparison of expected and actual effect. Within the scope of this project, I aim to simultaneously and systematically investigate both mechanisms of goal-directed action control - anticipatory action selection and proactive effect monitoring - as well as their interaction to gain a unified framework of action control. I will first examine proactive effect monitoring as reflected in anticipatory saccades more closely under different action control modes. Subsequently, I will systematically investigate which features of a future action effect are proactively monitored to determine the specificity of proactive effect monitoring. Finally, I will study the interaction of anticipatory action selection and proactive effect monitoring processes within a proposed unified model of human endogenous action control.