中文摘要：

采用线索化范式，通过实验任务设计线索与靶子不同的关联并鼓励被试采用不同的搜索策略，探讨线索与靶子间的知觉关联、语义关联及特征独子检测模式、特征搜索模式对注意捕获的作用。通过线索无效条件和线索有效条件反应时之差判断每一条件下的捕获量，并综合各条件的反应情况做出结论如下：（1）线索与靶子的知觉关联调控注意捕获，其效应不受语义关联和搜索策略影响；（2）线索与靶子的语义关联只在排除知觉关联后调控注意捕获；（3）控制知觉关联和语义关联后，特征搜索模式不影响注意捕获，而特征独子检测模式对注意捕获有影响。这表明语义知识和知觉经验根据当前任务要求共同作用指导个体的视觉加工，表现了人类加工系统的灵活性和可调节性。

英文摘要：

There is large amount of evidence showing that the saliency of a stimulus on a dimension （singleton） can involuntarily capture attention. Previous work has also demonstrated that attentional capture induced by feature singleton is modulated by top-down factors. For example, the semantic relationship between target and distractors impacts on the capture of attention; search strategies （Singleton Detection Mode and Feature Search Mode） are another potential factors affecting attentional capture. Here we measured how these factors guided attentional capture in a visual search task by manipulating cue-target relevance and search strategies and attempted to provide systematically evidence on capture of attention. A modified spatial cueing paradigm was employed in the current study. In a trial, a fixation screen was presented for 500 ms, followed by an uninformative red cue which appeared for 100 ms. Then a fixation screen was showed for 100 ms, followed by a target screen with Chinese characters surrounding by squares displayed 1000 ms. Participants had to make a judgment for the gap orientation of a target square while ignoring other distractors. Targets were Chinese characters randomly presented at the cued or non-cued locations, making cues valid or invalid. The effect of attentional capture referred to the fact slower responses to invalid contrasting to those to valid targets. The magnitude of the effect was computed to assess the effect size across different experimental conditions. In the current study, there were ten experimental conditions according to various combinations of perceptional relevance, semantic relevance between cue and target and singleton detection mode and feature search mode. The order of experimental conditions was counterbalanced across participants. As a result, the capture effect was not observed when cue and target were irrelevant whilst no search strategy was adopted, but feature search mode was induced. However, the capture effect was observed when singleton detection mode wa