Abstract
Significance: Both foveal and peripheral contour interaction are based on, as yet, unexplained neural mechanisms. Our results show that, unlike foveal contour interaction, peripheral contour interaction cannot be explained on the basis of the antagonistic structure of neural receptive fields.
Purpose: Foveal contour interaction is markedly reduced for mesopic compared to photopic targets. This finding is consistent with an explanation based on the antagonistic structure of neural receptive fields. However, no reduction was found for low-luminance targets in the periphery, possibly because the luminances used previously remained substantially above peripheral scotopic detection thresholds. In this study, we compared foveal and peripheral contour interaction for longwavelength photopic and mesopic targets, which would be expected to significantly elevate the peripheral retinal detection threshold.
Methods: Five normal observers viewed a randomly selected Sloan letter surrounded by four flanking bars at several edge-to-edge separations (min arc). Photopic and mesopic stimuli were viewed foveally and at 6 deg peripherally through a selective red filter that ensured mesopic targets were within 1 log unit of detection threshold at both retinal locations.
Results: Whereas the magnitude of foveal contour interaction was substantially less at mesopic compared to photopic luminance (20% vs. 46% reduction of percent correct, on average), no significant difference was observed in peripheral contour interaction, which had average mesopic and photopic magnitudes of 38% and 40%. Moreover, confusion matrices representing photopic and mesopic contour interaction differed in the fovea, but not in the periphery. The extent of contour interaction did not change with luminance at either retinal location.
Conclusions: Our results indicate that although the characteristics of foveal contour interaction can be accounted for by the antagonistic structure of neural receptive fields, the same mechanism is not compatible with the characteristics of peripheral contour interaction.
Purpose: Foveal contour interaction is markedly reduced for mesopic compared to photopic targets. This finding is consistent with an explanation based on the antagonistic structure of neural receptive fields. However, no reduction was found for low-luminance targets in the periphery, possibly because the luminances used previously remained substantially above peripheral scotopic detection thresholds. In this study, we compared foveal and peripheral contour interaction for longwavelength photopic and mesopic targets, which would be expected to significantly elevate the peripheral retinal detection threshold.
Methods: Five normal observers viewed a randomly selected Sloan letter surrounded by four flanking bars at several edge-to-edge separations (min arc). Photopic and mesopic stimuli were viewed foveally and at 6 deg peripherally through a selective red filter that ensured mesopic targets were within 1 log unit of detection threshold at both retinal locations.
Results: Whereas the magnitude of foveal contour interaction was substantially less at mesopic compared to photopic luminance (20% vs. 46% reduction of percent correct, on average), no significant difference was observed in peripheral contour interaction, which had average mesopic and photopic magnitudes of 38% and 40%. Moreover, confusion matrices representing photopic and mesopic contour interaction differed in the fovea, but not in the periphery. The extent of contour interaction did not change with luminance at either retinal location.
Conclusions: Our results indicate that although the characteristics of foveal contour interaction can be accounted for by the antagonistic structure of neural receptive fields, the same mechanism is not compatible with the characteristics of peripheral contour interaction.
Original language | English |
---|---|
Pages (from-to) | 1053-1060 |
Number of pages | 8 |
Journal | Optometry and Vision Science |
Volume | 97 |
Issue number | 12 |
Early online date | 25 Nov 2020 |
DOIs | |
Publication status | Published - 1 Dec 2020 |