Using Rapid Invisible Frequency Tagging to track internal attention

Abstract

When multiple items are held in Visual Working Memory (VWM), situational demands may require the prioritization of relevant ones. Previous studies measuring spatial attention using behavior, microsaccades, and electrophysiology suggest that locations where items were encoded play a role during their prioritization, even if their locations need not be explicitly remembered. Whether the engagement of spatial attention during prioritization also enhances sensory processing at encoding locations is currently unclear. Here we implemented a novel EEG paradigm using Rapid Invisible Frequency Tagging (RIFT) to quantify allocation of spatial attention during a VWM task. Participants encoded two items at distinct, frequency-tagged locations, one of which was cued retroactively. We quantified rhythmic brain responses to the individually tagged encoding locations over time as an index of attentional gain. Surprisingly, the RIFT response corresponding to the cued location did not increase following the retro-cue, even though we did observe a distinct cue-locked lateralization of alpha-oscillations - an established neural signature of spatial attention. In a second experiment we successfully verified that our RIFT setup was sensitive enough to measure attention directed to one of two perceived stimuli (as opposed to empty memory locations). Our results suggest that internal prioritization of memory items does not lead to increased visual processing at their encoding locations, despite the presence of other hallmarks of attention. We also show that attentional allocation can reliably be measured with RIFT using EEG (a far more inexpensive option than MEG). While our findings are compatible with previous studies, they reveal striking differences in the way that spatial attention modulates visual processing during the prioritization of memory, as compared to perceptual representations.