Abstract
Two recent studies of individual differences have provided new evidence that the stereoscopic disparity system represents depth structure via spatial-frequency channels (Peterzell et al., 2017, Vision Research; Reynaud & Hess, 2017, Frontiers. Comp. Neurosci.). Usually, it is assumed that these channels are tuned to retinal spatial frequency, in cycles per degree of visual angle (CPD). But could they show size-constancy, i.e. tuning to relative frequency, in ‘cycles per object’ (CPO)? (i.e. where “object” refers to the square aperture containing stimuli for both eyes, hence cycles relative to the object). First, we show that factors obtained by the two studies do not match when spatial frequency is specified in CPD. However, when the factors are replotted in CPO, the high frequency factor from the first study aligns with the low frequency factor from the second. Second, in a follow-up study, we measured thresholds for disparity-defined horizontal sinusoidal corrugations embedded in dots for 27 observers. CPO and CPD varied independently by changing object size: 0.3 cpd stimuli appeared at 1, 2, and 3 cpo; 1.0 cpd at 1, 2, 3, 6, 9 cpo; and 1.5 cpd at 1.5, 3, 6, 9, 12, and 15 cpo. We used an individual-difference analysis to identify factors. PCA of disparity sensitivities (log-arcsec), along with Varimax rotation replicated two factors from our 2017 study. Results are surprisingly inconclusive, with statistical factors/mechanisms seemingly jointly tuned to both retinal spatial frequency (CPD) and relative spatial frequency (CPO).