Abstract
At high temporal frequencies, early compressive nonlinearities within the visual system introduce artifacts enabling luminance (i.e. first-order) sensitive mechanisms to process contrast-modulated (i.e. second-order) stimuli. However, it is generally accepted that the impact of such artifacts can be canceled by adding a luminance grating. Furthermore, the fact that CM stimuli can be perceived even when compensating for such artifact is generally considered as evidence of two distinct motion processing pathways. The target of the present study was to investigate whether LM and CM processing at high temporal frequencies are common or distinct by trying to dissociate them using a noise masking paradigm. The CM stimuli were constructed so that the signal and carrier spatial frequencies importantly differed (0.5 cpd and 7-to-9 cpd, respectively). Using a direction discrimination task, contrast thresholds to LM (0.5 and 8 cpd) and CM (0.5 cpd) drifting (8 Hz) gratings were measured in dynamic noise filtered at either low (0.25-to-1 cpd) or high (4-to-16 cpd) spatial frequencies. For CM processing, early compressive nonlinearities were measured and compensated for by adding a LM grating. At the noise contrasts tested, the processing of 0.5 cpd LM gratings was more affected by low than high spatial frequency noise and the processing of 8 cpd LM gratings was more affected by high than low spatial frequency noise. This predictable double dissociation shows that the processing of 0.5 and 8 cpd LM gratings is, at least partially, distinct. CM processing was affected by similar proportions as luminance processing at 0.5 cpd. In other words, we failed to dissociate LM and CM processing (both at 0.5 cpd) suggesting that they are common. We propose a model within which non-uniform early nonlinearities across the visual field would cause non-uniform artifacts explaining CM processing at high temporal frequencies without inferring distinct processing.
This research was supported by NSERC operating grant to JF.