Abstract
Models of early cortical visual processing typically downplay the role of mean luminance in the neural coding of visual signals, emphasizing instead the coding of features such as relative contrast. Since the visual system is tasked with encoding surfaces and objects in scenes, which often vary independently in local luminance and contrast, it seems plausible that luminance information is encoded and plays an influential role in visuocortical processing. Indeed, electrophysiological studies in animals have found that increasing mean luminance levels multiplicatively increased the gain of contrast response functions in V1, most apparent at high contrast levels. In this study, we explored the degree to which different mean luminance levels can drive the early visual cortex in humans, using fMRI. We measured BOLD responses in early visual cortex (V1-V3) while participants viewed checkerboard stimuli that varied in contrast and mean luminance. Our experimental paradigm allowed us to reliably measure luminance response functions between 49 and 1278 cd/mm2 at two extreme contrast levels (4% and 96% Michelson Contrast), and at multiple spatial scales (voxel-wise and retinotopic). To control for changes in pupil diameter with varying luminance levels, stimuli were viewed monocularly through an artificial pupil. We found that luminance response functions in early visual cortex are contrast dependent. At high contrast, linearly increasing the mean luminance level produced a nonlinear increase in BOLD response. At low contrast, a flat BOLD response was observed across all mean luminance levels. These results reveal that the visuocortical neural code can represent information corresponding to changes in the mean luminance of a visual signal, and is most prominent at high contrast levels.
Meeting abstract presented at VSS 2018