Detection of flickering edges: absence of a red–green edge detector

Kelly ((1975) Science, 188, 371–372) showed that a centrally-fixated, contrast-reversing edge has a very different effect on the detection of luminance and red–green flicker. Red–green flicker sensitivity was ∼3-fold greater for a uniform field than for a ‘split’ field with the two sides flickering out-of-phase. Just the opposite effects were observed for luminance flicker — the split field yielded a 7-fold advantage over the uniform field at 2 or 4 Hz and a 3-fold advantage at 12 Hz. Contrary to Kelly, we find that the split field offers only a very small advantage of 40% for luminance flicker at 2 Hz and virtually no advantage at 4 Hz and above. Kelly’s chromatic results are surprising since one might expect that the larger color difference (or step) across the central edge would aid chromatic discrimination rather than strongly suppressing sensitivity. We show that the central chromatic edge only weakly impairs detection. Further results show that the two sides of the chromatic split field are detected essentially independently by red or green ‘blob’ detectors, which do not take advantage of the color difference across the edge. This has a remarkable implication: when wavelength discrimination is measured with a bipartite field whose two side are slowly modulated in opposite directions, then one side may be deleted with little adverse effect.