Abstract
Previously we used a gaze-contingent peripheral Gabor discrimination task to determine whether tunnel vision or general interference of attention results when (i) a foveal load is present and (ii) the peripheral stimuli have been scaled to be equally discriminable in single-task conditions (Ringer et al., 2016). We found that, while peripheral performance was near perfect and did not vary as a function of retinal eccentricity in single-task conditions, when simultaneously presented with a foveal L vs. T discrimination task, peripheral Gabor orientation sensitivity decreased with increasing eccentricity. The zoom-lens model of attention (Eriksen & Yeh, 1985; Eriksen & St. James, 1986) is the most intuitive explanation for these results, with a foveal attentional bias producing diminished peripheral sensitivity. However, our prior results do not account for potential competition between foveal and peripheral processing, and thus it is possible that multiple spotlights of attention (McMains & Somers, 2004) may explain processing peripheral and foveal targets in parallel. A zoom-lens account of attention predicts trade-offs between peripheral and foveal discrimination sensitivity, whereas a positive or neutral relationship between peripheral and foveal discrimination sensitivity would be predicted by a multifocal spotlight account. Multi-focal attention predicts processing difficulties as attention becomes increasingly divided, but not a foveal processing bias. Using our previous data, a hierarchical generalized linear mixed model was used to predict peripheral Gabor sensitivity as a function of retinal eccentricity, foveal response bias, and foveal task sensitivity. This showed peripheral Gabor sensitivity decreasing significantly with increased foveal load sensitivity. Though this effect did not interact with Gabor eccentricity, it is possible that our m-scaled peripheral stimuli mitigated this effect. Additionally, valid foveal stimuli (i.e., T present trials) produced faster, more accurate Gabor performance in the periphery, consistent with serial attentional switching. Thus we find converging evidence for the zoom-lens model of attention.
Meeting abstract presented at VSS 2017