We found evidence for the shifting trace model, an inversion of neural laterality signatures after the midline-crossing saccade. This was apparent in average MUA rate (Figure 4A) and in information carried by MUA (Figure 4B). As in the no-swap trials (Figures 2A and 2B), prefrontal MUA starts the delay with a bias toward the contralateral hemifield (Figures 4A and 4B, left; H symbols: p < 0.01, corrected, sample hemifield main effect in hemifield × shift condition permutation two-way ANOVA). But after the saccade (Figures 4A and 4B, right), both MUA and decoding accuracy increased for remembered locations that shifted from the ipsilateral to the contralateral hemifield (orange) relative to that for ipsilateral locations on no-swap trials (desaturated brown; brown stars: p < 0.01, paired t test). By contrast, on swap trials when the saccade shifted the remembered location from the contralateral to ipsilateral hemifield (green), there was a decrease compared with the contralateral location on no-swap trials (desaturated green; green stars: p < 0.01). For average MUA (Figure 4A), there was a complete inversion. The spike rates after a saccade
