Abstract
When looking at a familiar object, we can often imagine what it might feel like to the touch. Despite the multisensory nature of our interactions with common objects, working memory (WM) research has focused almost exclusively on unimodal stimuli. The current study investigated the recruitment of somatosensory representations during pure visual WM tasks as a function of haptic training with two classes of novel objects, Greebles and Fribbles. In a pilot experiment (N = 66), subjects completed a visual delayed match-to-sample (DMTS) task to confirm that memory performance for these novel objects was below ceiling (๐ = 0.77, ๐ = 0.07). Next, 7 subjects completed a pre-test on the visual DMTS task, followed by haptic and visual training, followed by a post-test on the same visual DMTS task. During training sessions, the objects were divided into three sets: one set for haptic training without visual feedback, one set for visual training without haptic feedback, and one โcontrolโ set of objects not trained. The results reveal improvement after training in both the visual condition and the cross-modal haptic condition, relative to the control condition. These results suggest that haptic training may improve performance even on a purely visual WM task, perhaps by recruiting additional representations of the remembered items in somatosensory cortex