Elliott et al. (2009) measured contrast sensitivity functions (CSFs) using both normal viewing and closed-loop adaptive optics (AO) to correct high-order optical aberrations. We reanalyzed their data to further characterize optical and neural contributions to CSFs. Data were log CSFs for 16 individuals measured monocularly with and without AO correction for six vertical sinewave gratings between 0.55 and 18 cycles per degree (cpd). A factor analysis (direct oblimin rotation) revealed three significant factors. The first loaded onto AO-corrected data for high spatial frequencies (1.25 to 18 cpd) but only onto 1.25 cpd for normal viewing. The second loaded onto low frequencies (0.55 to 2.25 cpd) regardless of viewing condition. The third loaded onto frequencies above 2.25 cpd, but only in normal viewing, and disappeared under AO correction. The first two factors were retino-neural, broadly sensitive to high and low spatial frequencies. The third factor was optical quality (variability in higher-order aberrations), indicating that previously identified factors above 2.25 cpd in factor-analytic studies are not retino-neural. Results may challenge the hypothesis of multiple narrowly tuned spatial frequency channels determining CSFs in favor of two broadly tuned neural mechanisms (consistent with magnocellular and parvocellular pathways) plus optical factors that limit high spatial frequencies. Notably, (typically older) observers with poorer optical quality exhibited reduced neural sensitivity, even after optical correction, suggesting chronic optical degradation may permanently impact neural processing.
Peterzell et al. (Mon,) studied this question.