How are different contrast effects on binocular luster spatially integrated?

In previous studies with dichoptic center-ring-surround stimuli, we found that two properties of the ring element have a strong influence on the phenomenon of binocular luster. The strength of the lustrous impression in the central target patch varies with increasing ring width, with the direction of this variation (increasing or decreasing) depending on the ring's luminance. In this study, we used stimuli in which the ring was split into segments with two different luminances that in uniform rings had opposite effects on perceived luster. The aim was to investigate how the lustrous impression is influenced by combining a weaker and a stronger contrast effect, in particular how they are spatially integrated by the visual system. In a psychophysical matching experiment, subjects had to assess the strength of the lustrous impression in a series of test stimuli with different ring widths, numbers of ring segments, and spatial proportions between the two ring parts. We found that the results of the experiment could neither be explained by a winner-takes-all integration (assuming that the lustrous response is completely determined by the stronger effect) nor by a balanced integration process (assuming equal weights for the two effects). Instead, both effects contribute to the overall lustrous response, with the stronger effect having a greater weight. Interestingly, the magnitude of this weight varied considerably between different groups of subjects. We found two main trends in the data, representing two different types of sensitivity to the phenomenon of binocular luster.