Abstract
The Contralateral Delayed Activity (CDA) has been a powerful tool with which to measure the representation of items in visual working memory. However, there are two reasons to believe that the CDA might reflect miniature gaze-shifts instead of working memory representations. First, to measure the CDA, subjects are presented with bilateral memory arrays but are asked to remember only items in one visual hemifield. This could elicit eye movements in the direction of the remembered items. Second, recent studies have shown that microsaccades modulate the occipitoparietal EEG and ERP responses (Yuval-Greenberg et al., 2008; Dimigen et al., 2009). To test the hypothesis that systematic microsaccades underlie the CDA, we recorded subjects’ EEG and averaged ERPs during change-detection tasks while also using high-resolution eye tracking (EyeLink 1000). Despite strict instructions to maintain fixation and the use of a stringent criterion for rejecting artifacts associated with eye movements, we found that the averaged gaze position was systematically shifted toward the hemifield of the memoranda during the retention interval. Moreover, the magnitude of these shifts increased with memory set size. However, we found that the CDA amplitude was insensitive to the magnitude of the gaze shifts within a given set size. We found an important dissociation between eye position measured with the video-based eye tracker and the horizontal electrooculogram (HEOG). The findings indicate that the HEOG picks up some of the neural activity underlying the CDA and the maintenance of representations in visual working memory. The findings demonstrate practical limitations in the use of HEOG to measure eye movements during ERP and EEG studies of visual memory. Additional implications of these results will be discussed in the context of the locus of the neural representations of working memory.
Meeting abstract presented at VSS 2013