People excel at learning the statistics of their environments. For instance, people rapidly learn to pay attention to locations that frequently contain visual search targets. Here, we investigated how frequently finding specific objects as search targets influences attentional selection during real-world object search. We investigated how learning that a specific object (e.g., a coat) is task-relevant affects searching for that object and whether a previously frequent target would influence search more broadly for all items of that target's category (e.g., all coats). Across five experiments, one or more objects from a single category were likely targets during a training phase, after which objects from many categories became equally likely to be targets in a neutral testing phase. Participants learned to find a single frequent target object faster than other objects (Experiment 1, N = 44). This learning was specific to that object, with no advantage in finding a novel category-matched object (Experiment 2, N = 32). In contrast, learning to prioritize multiple exemplars from one category spread to untrained objects from the same category, and this spread was comparable whether participants learned to find two, four, or six exemplars (Experiment 3, N = 72). These differences in the breadth of attention were due to variability in the learning environment and not differences in task (Experiment 4, N = 24). Finally, a set-size manipulation showed that learning affects attentional guidance itself, not only postselective processing (Experiment 5, N = 96). These experiments demonstrate that the breadth of attentional tuning is flexibly adjusted based on recent experience to effectively address task demands. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

中文翻译：

---

对对象类别的关注：选择历史决定了现实世界对象搜索期间注意力调整的广度。


人们擅长了解周围环境的统计数据。例如，人们很快学会注意经常包含视觉搜索目标的位置。在这里，我们研究了在现实世界的对象搜索过程中，寻找特定对象作为搜索目标的频率如何影响注意力选择。我们研究了了解特定对象（例如一件外套）与任务相关如何影响对该对象的搜索，以及先前频繁出现的目标是否会更广泛地影响对该目标类别的所有项目（例如所有外套）的搜索。在五个实验中，来自单个类别的一个或多个对象在训练阶段可能是目标，之后来自许多类别的对象在中性测试阶段同样可能成为目标。参与者学会了比其他对象更快地找到单个频繁目标对象（实验 1，N = 44）。这种学习是特定于该对象的，在找到新的类别匹配对象时没有优势（实验 2，N = 32）。相比之下，学习对一个类别中的多个样本进行优先级排序会传播到同一类别中未经训练的对象，并且无论参与者学会找到两个、四个还是六个样本，这种传播都是相当的（实验 3，N = 72）。这些注意力广度的差异是由于学习环境的变化而不是任务的差异造成的（实验 4，N = 24）。最后，设定大小的操作表明学习影响注意力引导本身，而不仅仅是后选择处理（实验 5，N = 96）。这些实验表明，注意力调节的广度可以根据最近的经验灵活调整，以有效满足任务需求。 （PsycInfo 数据库记录 (c) 2024 APA，保留所有权利）。