Abstract
The properties of S-cone flicker measured on longer-wavelength backgrounds are consistent with S-cone signals’ being transmitted by an achromatic pathway shared with L- and M-cone signals. One unusual property of this S-cone signal, which suggests that it is not a direct input, is that it is inverted in sign and delayed. We have looked in detail at S-cone temporal-frequency responses in normal observers and clinical observers with altered S-cone function. We confirm the findings of Wisowaty & Boynton [1980, Vision Research, 20, 895] that the S-cone responses with and without long-wavelength adaptation are different. On long-wavelength backgrounds, the data suggest the existence of multiple S-cone signals.
We model the results by supposing that the S-cones interact with L- and M-cones via a network of indirect lateral connections, and that each step in the network inverts and delays the S-cone signal. If the S-cone interaction is via one step, the signal is delayed and inverted, if it is via two steps it is doubly-delayed but "uninverted", and so on. The converging S-cone signals destructively or constructively interfere at different frequencies to produce characteristic patterns in the S-cone temporal frequency response. Our results further suggest that the relative strengths of 1-step, 2-step and 3-step S-cone signals vary between observers (presumably because of variability in the underlying S-cone photoreceptor density) to produce distinctive individual differences in the shapes of the S-cone temporal-frequency response. Our model is bolstered by measurements in observers with altered S-cone function, including an S-cone monochromat with only S-cones, enhanced-S-cone-syndrome patients with an excess of S-cones, and a CSNB patient without ON-pathways. The 1-, 2- and 3-step model makes specific predictions for each type of observer. We speculate that the network is provided by horizontal cells that feed back through L- and M-cones in single or multiple steps.
Meeting abstract presented at VSS 2013