The Attention Resource Required By The Maintenance Of Feature Binding In Visual Working Memory

Feature binding (FB) is a core concept in many cognitive fields (e.g. perception, working memory). During the past decade, it has been debated whether maintaining FB in visual working memory (VWM) is a more resource-demanding process than constituent features. Currently, it is generally considered a relatively automatic process without additional attention resource. However, previous studies have only explored the role of general attention or space-based attention in the FB maintaining process, which fail to examine the relationship between the maintenance of FB and object-based attention. Since FB processing is closely related to object representation in VWM, which may involve one special attention:object-based attention. The author hypothesized that maintaining FB in VWM require object-based attention resource, but not other attention resource.To test this hypothsis, the current study inserted a concurrent task to exhaust different type of attention resourse during the maintenance phase of VWM, and explored how maintenance of FB was impaired, that is whether the maintenance of FB was impaired more than constituent features (also called selectively impaired). The current study was composed of three parts:The first part aimed to examine the effect of consuming object-based attention resource on the maintenance of FB, by successively adopting mental rotation (Experiment 1), transparent motion (Experiment 2), and object feature report (Experiment 3) as the concurrent task to consume object-based attention resourse, and found out that the maintenances of FB were all selevtively impaired, and this result was extensive to three typical binding conditions:between different basic features, between basic feature and detailed feature, and between basic feature and location feature.The second part aimed to examine the effect of consuming space-based attention resource on the maintenance of FB, by respectively adopting visual search (Experiment 4) and backward counting/articular suppression (Experiment 5) as the concurrent task to consume space-based attention and general attention resoure, and found out that the maintenance of FB was not selectively impaired.The third part aimed to exclude some attention-free alternative explanations. included replacing the previous concurrent task with a visuo-spatial memory task to reject a representation-overwrite alternative (Experiment 6), extending the duration of maintenance phase to reject a memory-decay alternative (Experiment 7), and directly comparing the task load of the elected concurrent tasks to reject a task-difficulty alternative (Experiment 8).The current study mainly attained the following conclusion:The maintenance of feature binding in visual working memory require object-based attention resource, but not space-based or general attention resourse.