Abstract
This work investigated visual working memory (VWM) capacity load when directly manipulating its representations, as compared to just a maintenance process. Three experiments manipulated either visual addition, visual subtraction or mental rotation. Experiment 1 manipulated visual addition in which two shapes were presented and participants had to mentally ‘add’ them one on top of the other. Experiment 2 involved subtraction, such that the lower shape had to be subtracted from the upper shape, and in Experiment 3 participants mentally rotated letters that were presented in a canonical or mirrored form. Experiments 1 and 2 compared these online manipulations to just maintaining the same information, such that the visual information was identical across conditions. Experiment 3 included a control condition in which a canonical letter was presented in its upright position (no rotation). In Experiments 1 and 2, the results revealed a higher CDA amplitude when the items were added or subtracted, indicating a higher involvement of VWM, relative to just maintenance. In the mental rotation experiment, the results generally showed a higher CDA amplitude as the degree of rotation increased. Importantly, all experiments showed an increase in the CDA amplitude without an increase in the visual information presented. This increase in the CDA amplitude can be interpreted as indicating that the initial information together with the information that was created during the manipulation was represented in VWM. Another possibility could be that intermediate parts are being created while performing these manipulations, and maintaining the intermediate parts in VWM is responsible for the increase in the CDA amplitude.