Similar effects were seen in LFP power (Figure 5). During and just after the saccade, gamma (Figures 5A, 5B, and 5D) and theta (Figures 5A, 5B, and 5F) power were stronger overall for swap than no-swap trials. Later in the delay, however, gamma power inverted and became stronger for remembered locations moving from the ipsilateral to the preferred contralateral hemifield (Figures 5C and 5D; p < 0.01, hemifield × swap interaction effect) as predicted by the shifting trace model. Theta power showed a similar inversion at the very end of the delay, though this is likely due to temporal smearing of test period effects (Figures 5C and 5F; gray regions indicate extent of possible test effects).
