Perception-memory interactions reveal a computational strategy for perceptual constancy.

Area of Science:

• Visual perception
• Cognitive neuroscience
• Psychophysics

Background:

• The visual system faces challenges in extracting constant object properties due to ambiguous and noisy sensory information.
• Perceptual constancy of surface lightness is explained by two main model classes: contrast-based and reflectance-inference models.

Purpose of the Study:

• To test competing models of lightness perception using a novel psychophysical task.
• To investigate the effects of memory demand and context change on lightness perception.
• To determine whether lightness perception relies on contrast coding or reflectance inference.

Main Methods:

• Developed a novel psychophysical task incorporating memory demand (stimulus delay) and context change (surround luminance variation).
• Implemented probabilistic versions of contrast-based and reflectance-inference models.
• Analyzed observer performance under different experimental conditions, including combined memory and context manipulations.

Main Results:

• Memory demand biased perceived lightness towards the ensemble mean, consistent with prior expectations.
• Context change induced the simultaneous lightness contrast effect.
• The interaction between memory and context was less than predicted by independent effects, and the reflectance model better explained the magnitude and interaction of these effects.

Conclusions:

• Lightness perception is better explained by reflectance inference than by contrast coding.
• Incorporating memory load into perceptual tasks enhances ecological validity and aids in distinguishing between competing perceptual models.