Characterizing the influence of spatial attention on stimulus-evoked cortical representations

Abstract

Past work suggests that covert attention shifts spatial receptive fields (RFs) towards attended locations, increasing the neuronal resources devoted to representing the attended location (Anton-Erxleben & Carrasco, 2013). If such shifts of spatial RFs occur on a large scale, the size of the cortical representation of visual stimuli should expand because attended stimuli will fall within the RFs of a broader range of neurons. Past work with fMRI, however, has found increases in the amplitude of the stimulus representations, but no substantial change in their size (Sprague & Serences, 2013; Võ et al., 2017), a finding that is hard to reconcile with the hypothesis that attention shifts spatial RFs towards attended locations. Two factors motivate further examination of this question. First, prior studies have shown that increases in BOLD activity at attended locations are additive (i.e., independent) of stimulus contrast, suggesting that this measure may not tap into modulations of stimulus-evoked responses (Murray, 2008). Second, direct comparisons of fMRI and EEG activity have revealed qualitatively distinct modulations by attention (Itthipuripat et al., 2018). Thus, we re-examined this question with a focus on stimulus-evoked EEG activity. We measured the response evoked by attended and unattended stimuli. Using an IEM, we reconstructed spatial channel-tuning functions (CTFs) from the broadband EEG response evoked by the stimuli. This analysis revealed that both the amplitude and the size of the stimulus-evoked CTF was larger for attended than for unattended stimuli. The change in the size of the spatial CTFs cannot be explained by sensory gain alone, which can only change the amplitude of CTFs. Thus, our results suggest that covert spatial attention shifts RFs towards the attended location, expanding the cortical representation evoked by stimuli in that location.