Different internal noise but same calculation efficiency for processing luminance-modulated (LM) and contrast-modulated (CM) stimuli

Abstract

Introduction: We are generally less sensitive to CM than LM stimuli. Therefore, CM stimuli are typically considered more complex than LM stimuli. Why are we less sensitive to CM stimuli when an ideal observer is not? Based on known theoretical formulations, measuring the sensitivity of a stimulus in different noise conditions permits to indirectly evaluate the subject's internal equivalent noise and calculation efficiency for perceptual processing of that stimulus. The purpose of the study was to compare internal equivalent noise and calculation efficiency for LM and CM stimuli processing in order to identify the level where the perceptual processing differs. Methods: Both LM and CM stimuli were composed of a 1 cycle/deg sinusoidal signal and a plaids (two perpendicular 8 cycles/deg sinusoidal gratings). The luminance profiles of the LM and CM stimuli were given by the sum and multiplication, respectively, of these two components. The detection thresholds of five young subjects were evaluated for both types of stimulus in three different noise conditions. A regression analysis was performed on the data to derive each subject's internal equivalent noise and calculation efficiency for LM and CM stimuli perception. Results: The internal equivalent noise was greater for CM than LM perceptual processing. However, the calculation efficiencies were not significantly different. Conclusion: The present study suggests that a relatively low-level processing difference between the detection of LM and CM stimuli explains the lower CM sensitivity. A model was proposed where a second-order rectification process would be sub-optimal and where higher-level perceptual mechanisms would be common for both LM and CM stimuli processing.