Abstract
Working memory (WM) stores detailed representations in an online state by recruiting the same cortical regions that encoded those items (e.g., Serences et al., 2009; Harrison & Tong, 2009). More recently, we found that mnemonic precision is correlated with the dispersion of orientation-specific population responses in primary visual cortex, indicating that these population codes play a functional role in supporting the fidelity of WM representations. In the current work, we investigated the mechanism by which these population responses are knit together across distinct regions of visual cortex. One prominent hypothesis is that activity across disparate neural units is bound together by virtue of temporal synchronization (Fries, 2005; Fell & Axmacher, 2011). Here, we corroborate this hypothesis with evidence that the degree of phase synchronization across brain regions strongly predicts individual differences in mnemonic precision. Observers stored a single foveally presented orientation in visual working memory while we measured EEG. Neural synchrony was measured by examining the degree to which the phase of ongoing oscillatory activity was matched across an electrode pair. The key finding was that phase synchronization (in the upper alpha to lower beta bands, 10-15 Hz) was strongly predictive of which observers had the most precise recall of the stored orientation; this was observed across both occipital (R[sup]2[/sup]=.3) and parietal (R[sup]2[/sup]=.33) electrode pairs. Finally, we collected the same measures using a lateralized stimulus and found that (1) The phase of oscillatory activity measured in the ipsilateral electrode lagged behind that in the contralateral electrode until 200 ms after the onset of the stimulus, and (2) The relationship between WM precision and phase synchrony was observed only after interhemispheric synchrony was established. Together, these studies support the hypothesis that the precision of visual WM representations is determined by the sustained global synchronization of feature-selective population codes.
Meeting abstract presented at VSS 2013