How can humans adaptively use temporal regularities within their environment to help guide attention?

现实世界的环境不是完全随机的，而是充满规律的。我们可以利用这些规律来形成对未来相关事件的预测，并主动指导我们的行为。使用静态视觉搜索任务（参与者在其他分散注意力的物品中搜索目标）的研究表明，我们可以在繁忙的视觉场景中学习规律。此外，我们可以使用这些规律来预测目标会出现在哪里，或者它们会是什么样子，并相应地将我们的注意力引导到这些位置的物品上，或者利用这些视觉特征，更有效地找到目标。然而，与这些任务中的视觉搜索数组不同，现实世界的场景是动态的。因此，必须考虑时间和时间规律对视觉搜索的影响。时间规律对搜索的影响直到最近才开始被研究。最初的研究是通过一个动态视觉搜索任务来完成的，在这个任务中，某些目标可以预见地出现在同一地点和同一时间。其他目标和干扰物出现在不可预测的地点和时间。在这里，参与者识别时空可预测目标的频率和速度明显高于时空不可预测目标。这些发现表明，参与者学习了时间规律，这里结合了空间规律，并利用这些知识将他们的注意力引导到预期目标出现的特定位置。这一初步证据表明，在视觉搜索过程中，人类可以将基于规则的时间预测作为一种有用的注意力引导来源。然而，我们对可能发生这种情况的情况知之甚少，也不知道这种依赖时间的指导的性质。这个项目旨在为这些问题提供深入的见解。例如，我将研究我们是否可以使用时间规律来指导搜索，当它们独立于空间位置运行时，并结合非空间特征，如颜色。此外，我对参与者在搜索任务中的运动反应如何支撑他们对时间规则的学习和使用感兴趣。我将使用动态视觉搜索任务来研究这些问题和其他问题，在这些任务中，目标和干扰物在试验期间的不同时间出现，有时是可预测的。根据手头的问题，这些任务的要素，如目标的可预测属性（例如，位置和/或颜色和/或时间），或任务所需的运动反应，会有所不同。我还将开发新的方法来获取和分析时间序列数据，在人们执行动态视觉搜索任务时捕捉连续的行为和大脑测量。与典型的准确性或反应时间测量相比，这些测量将更有助于理解基于规则的时间预测如何准确地随着时间的推移影响注意力指导。例如，我将测量参与者在一段时间内用眼睛盯着特定干扰物的可能性。我将评估在目标与干扰物共享特征可预测出现之前，这种可能性是如何变化的。干扰物的注视可能性的增加反映了目标特征的注意优先级的增加。在这一点上，看看在一个暂时可预测的目标出现之前多久，这种增长可能会开始，这将是特别有趣的。总的来说，这个项目旨在扩大和加深我们对人类如何在环境中学习和使用时间规律的理解，以形成预测，并随着时间的推移自适应地引导他们的视觉注意力。这个项目将有助于我们理解如何在繁忙和动态的现实世界中高效地执行任务。此外，它将有助于更广泛的研究，调查选择性视觉注意中以前被忽视的时间主题。

Real-world environments are not entirely random but imbued with regularities. We can use these regularities to form predictions about future relevant events and proactively guide our behaviour. Research using static visual-search tasks, where participants search for targets amongst other, distracting items, has shown that we can learn regularities within busy visual scenes. Further, we can use these regularities to predict where targets will appear, or what they will look like, and accordingly guide our attention towards items in these locations, or with these visual features, to find targets more efficiently. However, unlike the visual-search arrays in these tasks, real-world scenes are dynamic. Thus, it is imperative to consider the effects of time, and temporal regularities, on visual search. Effects of temporal regularities on search have only recently begun to be investigated. Initial studies have done so using a dynamic visual-search task where certain targets predictably appeared in the same location and at the same time during trials. Other targets, and distractors, appeared at unpredictable locations and times. Here, participants identified spatiotemporally predictable targets significantly more often, and significantly faster, than spatiotemporally unpredictable targets. These findings suggest that participants learnt temporal regularities, here combined with spatial regularities, and used this knowledge to guide their attention towards certain locations at times when targets were expected to appear there. This initial evidence suggests that humans can use regularity-based temporal predictions as a helpful source of attentional guidance during visual search. However, we know little about the circumstances under which this can occur, or the nature of this time-dependent guidance. This project aims to provide insight into these questions. For example, I will investigate whether we can use temporal regularities to guide search when they operate independently of spatial location, and in combination with non-spatial features like colour. Further, I am interested in how participants' motor responses during search tasks may scaffold their learning and use of temporal regularities. I will investigate these questions, and others, using dynamic visual-search tasks in which targets and distractors appear at different times during trials, sometimes predictably. Depending on the question at hand, elements of these tasks, such as the predictable properties of targets (e.g., location and/or colour and/or time), or the motor responses the tasks require, will vary. I will also develop new methods to acquire and analyse time series data capturing continuous behavioural and brain measures while people perform the dynamic visual-search tasks. These measures will be more informative for understanding how exactly regularity-based temporal predictions may shape attentional guidance over time, compared to typical accuracy or response-time measures. For example, I will measure the likelihood of participants fixating, with their eyes, a particular distractor over time. I will assess how this likelihood changes in the moments before a target sharing features with that distractor predictably appears. An increase in fixation likelihood of the distractor here would reflect an increase in attentional prioritisation of target features. Here, it will be particularly interesting to see how soon before a temporally predictable target appears this increase may begin. Overall, this project aims to expand and deepen our understanding of how humans can learn and use temporal regularities in their environment to form predictions and adaptively guide their visual attention over time. This project will contribute to our understanding of how we can perform tasks efficiently and effectively in the busy and dynamic real world. Further, it will contribute to wider research investigating the previously neglected topic of timing in selective visual attention.