Abstract
Even though neurons in early visual areas respond in a contrast dependent manner, we can consciously perceive a stimulus even at relatively low contrast. To test the hypothesis that the precision of these high-level representations remains relatively constant at low levels of contrast, we conducted two experiments. The first was a psychophysical experiment to demonstrate that at low contrast people could have a precise representation of an oriented-teardrop-shape. Teardrop-shaped stimuli were presented as dark stimuli on a grey background. One stimulus was black (86% contrast), while the other two stimuli were only slightly darker than the background (.8%, 1.8% contrast). Participants were instructed to remember the orientation of the teardrop and to reproduce it after a 1500ms delay. We found that as the contrast level was reduced the precision was unaffected, although at the lowest contrast we saw an increase in the guess rate. We then performed an ERP decoding experiment where we examined the ability to decode the orientation of the stimulus from the scalp distribution of the ERP at each moment in time following the onset of the stimulus. This task was similar to the behavioral task, with slight modifications. This allowed us to test the hypothesis that early sensory activity would be reduced and therefore undecodable at low contrast, whereas later, higher-level processing (presumably conscious) would be decodable in both high and low contrast conditions. Results were consistent with our hypothesis, showing that during the early sensory period (~50-150ms) decoding accuracy was higher for high contrast than low contrast. By 200ms, decoding accuracy was equally high for low contrast stimuli as for high contrast stimuli. These findings suggest that early visual areas are not encoding the conscious perception of orientation, but awareness instead requires iterative processing over time to extract a largely contrast independent perception of orientation.