Abstract
[Background] Visual performance is better along the horizontal meridian (HM) than the vertical meridian (VM) and better at the lower than upper VM. However, the computations underlying these behavioral asymmetries are unknown. Here, we investigated (1) how the visual system differentially represents two fundamental visual features–orientation and spatial frequency (SF)–around the visual field and (2) whether behavioral asymmetries manifest as differences in featural representation. [Method] Observers detected a 2-cpd horizontal Gabor embedded in noise appearing at the fovea or at 4 cardinal locations at 6° eccentricity. We calculated contrast sensitivity by taking the reciprocal of the Gabor contrast titrated per location. We used reverse correlation to estimate the weights assigned by the visual system to a range of orientations (-90 to 90°) and SFs (1 to 4 cpd), conventionally interpreted as perceptual sensitivity to the corresponding feature. Then, tuning functions were derived from weight profiles and compared across locations. Last, we assessed the correlation between performance and tuning functions. [Results] We found (1) higher contrast sensitivity at the HM than VM and at the lower than upper VM; (2) higher baseline sensitivity to orientation at the HM than VM but no difference along the VM; (3) higher perceptual sensitivity to SFs at the HM than VM and at the lower than upper VM, especially to task-relevant SFs at the HM than VM with no difference in the bandwidth; (4) a positive correlation between the contrast sensitivity and the perceptual sensitivity to task-relevant orientations and SFs across observers. [Conclusion] Results suggest that a better representation of visual features underlies observed behavioral advantages at the HM and lower VM. Specifically, better detection at the HM than VM and at the lower than upper VM could be attributed to higher perceptual sensitivity to task-relevant features.