Electrophysiology reveals distinct neural mechanisms for lateralized and spatially global visual working memory representations.

Visual working memory (VWM) allows us to actively maintain a limited amount of information. To study the neural basis of its limited capacity, electrophysiological studies have traditionally employed a task in which participants remember items presented in a cued hemifield, while ignoring the items in the other hemifield. Such studies have successfully identified lateralized neural correlates of VWM capacity by measuring a sustained negativity over the contralateral hemisphere to the cued hemifield relative to the ipsilateral hemisphere (e.g., the contralateral delay activity, or the CDA, Vogel & Machizawa, 2004). Another line of work measured lateralized alpha power modulations and proposed that these oscillations may underlie the CDA and measure this same lateralized VWM representation (Sauseng et al, 2009; Mazehari & Jensen, 2008). However, recent fMRI studies have reported that VWM representations can be decoded not only from the contralateral hemisphere, but also from ipsilateral hemisphere (Ester, Serences, & Awh, 2011; Pratte & Tong, 2014). To test whether this spatially global spread of the VWM representations can be measured electrophysiologically, we ran two experiments in which we independently manipulated the number of targets and distractors. We found that irrespective of the number of distractors, the CDA tracked the number of items held in VWM, thus indexing the contralateral VWM representations. In contrast, the parieto-occipital alpha power over both hemispheres tracked the number of items held in VWM, thus indexing the spatially global VWM representations. These findings are consistent with the previous fMRI findings indicating the existence of spatially global VWM representations. In addition, our findings suggest the event-related potentials and alpha-band oscillations index different neural mechanisms that appear to map onto lateralized and spatially global representations, respectively. Meeting abstract presented at VSS 2015.