Probabilistic rejection templates in visual working memory

Our interactions with the visual world are guided by attention and visual working memory. Things that we look for and those we ignore are stored as templates that reflect our goals and the tasks at hand. The nature of such templates has been widely debated. A recent proposal is that these templates can most straightforwardly be thought of as probabilistic representations of task-relevant features. We assessed observers’ representations by measuring the slowing of visual search when distractor templates unexpectedly match the target. To provide a strong test of the templates’ probabilistic nature, distractor stimuli were heterogeneous, randomly drawn on each trial from a bimodal probability distribution. Using two targets on each trial, we tested whether observers encode the full distribution, only one peak of it, or the average of the two peaks. Search was slower when the two targets corresponded to the two modes of previous distractor distributions than when one target was at one of the modes and another between the modes or outside the distribution range. Furthermore, targets on the modes were reported later than targets between the modes that, in turn, were reported later than targets outside this range. These results show that observers represent both distribution modes using templates based on probability distributions rather than single features or simple summary statistics. Our findings indicate that visual working memory templates guiding attention are probabilistic and dynamically adapt to task requirements, reflecting the probabilistic nature of the input.

我们的视觉世界互动受注意力和视觉工作记忆的引导。我们所寻找的事物与我们所忽略的事物均被存储为模板，这些模板反映了我们的目标和当前任务。此类模板的本质已被广泛讨论。近期的一项提议认为，这些模板最直接的思考方式是作为任务相关特征的概率表示。我们通过测量当干扰模板意外地与目标匹配时视觉搜索的减速来评估观察者的表征。为了对模板的概率性质进行强有力的测试，干扰刺激在每次试验中均从双模态概率分布中异质地随机抽取。使用每次试验中的两个目标，我们测试了观察者是否编码了整个分布、其一个峰值或两个峰值的平均值。当两个目标对应于先前干扰分布的两个模式时，搜索速度比一个目标位于一个模式而另一个位于两个模式之间或分布范围之外时慢。此外，位于模式上的目标比位于模式之间的目标报告得晚，而位于模式之间的目标又比位于此范围之外的目标报告得晚。这些结果表明，观察者使用基于概率分布的模板，而不是单一特征或简单的汇总统计数据，来表征分布模式。我们的发现表明，引导注意力的视觉工作记忆模板是概率性的，并能动态适应任务要求，反映了输入的概率性质。