Modeling visual performance differences ‘around’ the visual field: A computational observer approach

Kupers, Eline R; Carrasco, Marisa; Winawer, Jonathan

doi:10.1371/journal.pcbi.1007063

Cited by 51 publications

(74 citation statements)

References 93 publications

(170 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cone density becomes sparser with eccentricity, due to increased size and larger gaps between cones, and decreases by ~30% between the HM and VM at a fixed eccentricity Song et al, 2011). A computational observer model has been used to evaluate the extent to which these optical and retinal factors can explain performance differences in contrast sensitivity with polar angle (Kupers, Carrasco & Winawer, 2019). To account for the 30% increase in contrast sensitivity between the UVM and the HM for stimuli (4 cpd) presented at 4.5º eccentricity (Cameron et al, 2002), the model required an increase by ~7 diopters of defocus or a reduction by 500% in cone density, which exceed by far the variations observed in human eyes.…”

Section: Discussionmentioning

confidence: 99%

Asymmetries in visual acuity around the visual field

Barbot¹,

Xue²,

Carrasco³

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Human vision is heterogeneous across the visual field. At a fixed eccentricity, performance is better along the horizontal than the vertical meridian, and better along the lower than the upper vertical meridian. These asymmetric patterns, termed performance fields, have been found in numerous visual tasks, including those mediated by contrast sensitivity and spatial resolution. However, it is unknown whether spatial resolution asymmetries are confined to the cardinal meridia or whether, and how far, they extend into the upper and lower hemifields. Here, we measured the extent of the vertical meridian asymmetry by assessing spatial frequency (SF) sensitivity at isoeccentric peripheral locations (10 degree eccentricity), every 15º of polar angle. On each trial, observers judged the orientation (±45º) of a suprathreshold grating stimulus varying in spatial frequency. We estimated 75%-correct SF thresholds, SF cutoff points (i.e., chance-level) and slopes of the psychometric function for each of the 24 isoeccentric locations. We found higher SF sensitivity in the lower than the upper hemifield, with this asymmetry being most pronounced at the vertical meridian and decreasing gradually as the angular distance from the vertical meridian increased. This gradual change in SF sensitivity with polar angle reflected a shift of the psychometric function without reliable changes in slope. The same pattern was found under binocular and monocular conditions. These findings advance our understanding of visual processing around the visual field and help constrain models of visual perception.

show abstract

Section: Discussionmentioning

confidence: 99%

Asymmetries in visual acuity around the visual field

Barbot¹,

Xue²,

Carrasco³

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, the recent implementation of a computational observer model has identified that asymmetries in cone sampling and optics can only account for a small portion of isoeccentric differences in contrast sensitivity. Specifically, the model found that a ~30% decrease in contrast sensitivity between the HM and UVM would require a 500% decrease in cone density, which is well beyond the actual density change (Kupers, Carrasco, & Winawer, 2019). Notably, cone densities can vary up by to a factor of 3 across individuals (Curcio, Sloan, Packer, Hendrickson, & Kalina, 1987).…”

Section: Neural Substrates Of Performance Fields Asymmetriesmentioning

confidence: 97%

Stimulus-dependent contrast sensitivity asymmetries around the visual field

Himmelberg

Winawer

Carrasco

2020

Preprint

Self Cite

View full text Add to dashboard Cite

AbstractAsymmetries in visual performance at isoeccentric locations are known as performance fields. At a fixed eccentricity, visual performance is best along the horizontal, intermediate along the lower vertical, and poorest along the upper vertical meridian. These performance fields are pervasive across a range of visual tasks, including those mediated by contrast sensitivity. However, contrast performance fields have not been characterized with a systematic manipulation of stimulus spatial frequency, eccentricity, and size; three parameters that constrain contrast sensitivity. Further, individual differences in performance fields measurements have not been assessed. Here, we use an orientation discrimination task to characterize the pattern of contrast sensitivity across four isoeccentric locations along the cardinal meridians, and to examine whether and how this asymmetry pattern changes with systematic manipulation of stimulus spatial frequency (4 cpd to 8 cpd), eccentricity (4.5° to 9°), and size (3° visual angle to 6° visual angle). Our data demonstrate that contrast sensitivity is highest along the horizontal, intermediate along the lower vertical, and poorest along the upper vertical meridian. This pattern is consistent across stimulus parameter manipulations, even though they cause profound shifts in contrast sensitivity. Eccentricity-dependent decreases in contrast sensitivity can be compensated for by scaling stimulus size alone. Moreover, we find that individual variability in the strength of performance field asymmetries is consistent across conditions. This study is the first to systematically and jointly manipulate, and compare, contrast performance fields across spatial frequency, eccentricity, and size, and to address individual variability in performance fields.

show abstract

“…For the main analysis, we use a linear support vector machine (SVM) classifier to discriminate if a stimulus orientation is clockwise or counter-clockwise from vertical given the RGC responses. As in our previous model [46], we apply a 2D fast Fourier transform to the RGC responses at each time point, and remove phase information. Classifying the mRGC amplitudes, rather than the full Fourier representation, allows the inference engine to learn a stimulus template that is robust to small eye movements and stimulus phase, and to categorize the stimulus orientation (illustrated by the high values at the stimulus spatial frequency ( Fig 3 , panel 5).…”

Section: Resultsmentioning

confidence: 99%

“…For comparison with our observer model based on cone outputs in our prior work [46], we used the same simulations as previously (identical eye movement patterns, cone densities, L-, M-, S-cone distributions), and to each of these simulations we added one of five mRGC layers, varying in density and filter widths.…”

Section: Resultsmentioning

confidence: 99%

Asymmetries around the visual field: From retina to cortex to behavior

Kupers

Benson

Carrasco

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Visual performance varies around the visual field. It is best near the fovea compared to the periphery, and at iso-eccentric locations it is best on the horizontal, intermediate on the lower, and poorest on the upper meridian. The fovea-to-periphery performance decline is linked to the decreases in cone density, retinal ganglion cell (RGC) density, and V1 cortical magnification factor (CMF) as eccentricity increases. The origins of radial asymmetries are not well understood. Optical quality and cone density vary across the retina, but recent computational modeling has shown that these factors can only account for a small percentage of behavior. Here, we investigate how visual processing beyond the cones contributes to radial asymmetries in performance. First, we quantify the extent of asymmetries in cone density, midget RGC density, and V1-V2 CMF. We find that both radial asymmetries and eccentricity gradients are amplified from cones to mRGCs, and from mRGCs to cortex. Second, we extend our previously published computational observer model to quantify the contribution of spatial filtering by mRGCs to behavioral asymmetries. Starting with photons emitted by a visual display, the model simulates the effect of human optics, fixational eye movements, cone isomerizations and mRGC spatial filtering. The model performs a forced choice orientation discrimination task on mRGC responses using a linear support vector machine classifier. The model shows radial asymmetries in performance that are larger than those from a model operating on the cone outputs, but considerably smaller than those observed from human performance. We conclude that spatial filtering properties of mRGCs contribute to radial performance differences, but that a full account of these differences will entail a large contribution from cortical representations.

show abstract

Modeling visual performance differences ‘around’ the visual field: A computational observer approach

Cited by 51 publications

References 93 publications

Asymmetries in visual acuity around the visual field

Asymmetries in visual acuity around the visual field

Stimulus-dependent contrast sensitivity asymmetries around the visual field

Asymmetries around the visual field: From retina to cortex to behavior

Contact Info

Product

Resources

About