Abstract
Several methods have been used to determine the relative modulation of the L and M cones that minimizes luminance modulation, including the minimally-distinct border (MDB; Boynton, 1978), heterochromatic flicker photometry (HFP; Ives, 1912), and minimum motion (MM; Anstis & Cavanagh, 1983) techniques. When performed on a modern computer-driven display, MDB is the simplest to set up (code, calibrate, and run), MM is likely to be the easiest for naive observers to learn and perform, and HFP is intermediate on both of these dimensions. We compared the three techniques in terms of their variability across measurements, their agreement with one another, and their test-retest reliability. In all techniques, the ~2 deg stimuli were presented at the center of a uniform grey background (59 cd/m2). Observers used buttons to vary the stimulus angle in the (L, M) plane of cone contrast space – S cones were not modulated -- either at 25% or 50% of the maximum contrast available at each angle, to find the equilu-minant angle. In HFP, observers minimized apparent flicker of a square at 10.75 Hz. In MM, observers minimized the motion of a 2.5 cpd counterphase grating at 2 Hz. In MDB, observers attempted to minimize the distinctness of an ‘isolated edge’ (a horizontal step multiplied by a Gaussian; McLellan et al., 1994) in two variants: 500 ms flashes (MDB1), or steady presentation (MDB2). A second session was done at least a week after the first session. At all mean equiluminant settings, L and M contrasts were of opposite sign, as expected. Across observers, HFP produced the lowest |L/M| ratio (smallest equiluminant angle) while MDB1 and MDB2 gave somewhat larger |L/M| ratios. Inter-method correlation coefficients showed only moderate consistency across methods. MM and HFP showed high test-retest reliability while MDB1 did not. Additional comparisons will be discussed.
Acknowledgement: NIH Grant R01 EY023281