Rose Nasrawi¹ (), Freek van Ede¹; ¹Institute for Brain and Behavior Amsterdam, Department of Experimental and Applied Psychology, Vrije Universiteit Amsterdam

Visual information maintained in working memory can be used to plan for and guide upcoming behavior. Yet, in daily life, the order in which visual information is encoded into working memory does not always correspond with the order in which this information will become relevant for behavior: visual information that is encoded early might only become relevant for behavior later, and vice versa. We asked how the dynamics of visual selection and planning for prospective memory-guided behavior depend on the order in which visual information is anticipated to become relevant for behavior. To address this, we developed a visual-motor working-memory task in which participants were asked to sequentially encode and memorize two colored tilted bars, that were reproduced consecutively following two separate response cues. The order in which the items needed to be reproduced could always be anticipated by the participant, and varied between trials: the cueing-order could either be the same as the encoding-order (first item cued for report first), or be reversed (first item cued for report second). We used EEG time-frequency analyses to track the selection of visual information (item location) and action planning (prospective response hand, linked to item tilt) during working memory. We show that visual selection of both items is largely unaffected by the order in which they are expected to become relevant for behavior. In contrast, signatures of action planning exhibit a clear prioritization of the action that will become relevant for behavior first. Furthermore, they show how planning to act on one memory item can take place alongside visual encoding of another item. These results reveal that visual encoding and planning for prospective manual actions can be decoupled, whereby the brain can flexibly encode new visual information into working memory, while concurrently planning to guide behavior using previously encoded visual information.

Acknowledgements: This research was supported by an ERC Starting Grant from the European Research Council (MEMTICIPATION, 850636) and an NWO Vidi grant by the Dutch Research Council (grant number 14721)