Abstract
To unravel the connection between color perception and individual photoreceptors, two key hurdles need to be overcome. First, objectively identifying the cone types in the trichromatic cone mosaic is required to characterize their spatial and spectral topography. Second, recording color sensations upon controlled, fine-grained, spatio-temporal stimulation of cones is needed to drive specific retinal circuits in isolation. To date, only full-field adaptive optics (AO) fundus photography has been used to map the human trichromatic cone mosaic. First, we sought to investigate the feasibility of employing an AO scanning laser ophthalmoscope (AOSLO) for objective spectral classification. AO imaging and retinal densitometry was conducted to classify S-, M- and L-cones following methods proposed by Roorda & Williams, 1999. The average probability of identifying L & M cones was 0.97, while the uncertainty in their identity was ~4%. The advantages obtained in resolution, contrast and video rate imaging via the AOSLO facilitated cone classification within 10-15 bleaching cycles. Next, in a preliminary study, cone-targeted color sensations were measured on a spectrally classified retina at 1.5 deg eccentricity. Cones were individually stimulated on a white background with AO-corrected and retinally stabilized stimuli at 543nm. A 2-dimensional hue-saturation color space was used to record responses. The subject required only red, green, white and ‘not seen’ as the categories for his responses. The consistency of reported color on average across cones was 70%. Blue or yellow percepts were never reported. Overall, white was reported about 50% of the time across all cones. S-cones did not produce a detectable percept (‘not seen’) in about 50% of trials. In cases when a color other than white was reported, L-cones signaled ‘red’ while M-cones signaled ‘green’. No opposing colors were reported in either L or M cones. Although these first results indicate that the hue reporting cones simply follow the trichromatic color theory, altering the stimulation and background wavelengths should allow probing opponent retinal circuitry in the future.