Characterizing ensemble statistics: mean size is represented across multiple frames of reference.

Area of Science:

• Visual perception
• Cognitive neuroscience
• Computational vision

Background:

• The human visual system often perceives statistical properties of visual sets rather than individual elements.
• Understanding how the brain represents ensemble statistics in different spatial frames of reference is crucial for explaining visual perception.

Purpose of the Study:

• To investigate the spatial nature of ensemble statistics, specifically how average size is encoded across multiple reference frames.
• To determine if the visual system's representation of mean size is retinotopic, spatiotopic, or hemispheric.

Main Methods:

• Employed a mean-size adaptation paradigm where observers were adapted to patches of small and large dots in opposing display regions.
• Tested perception of single dots presented in retinotopic, spatiotopic, and hemispheric coordinates relative to adaptation areas.
• Examined cross-eyed transfer of adaptation aftereffects.

Main Results:

• Observed significant retinotopic, spatiotopic, and hemispheric adaptation aftereffects for perceived size.
• Participants perceived test dots as larger in areas previously adapted to smaller dot ensembles.
• Adaptation aftereffects demonstrated transfer between the observer's eyes.

Conclusions:

• Mean size is represented and processed across multiple spatial frames of reference within the visual system.
• These findings support the role of ensemble statistics in maintaining stable visual perception despite changes in individual elements.
• The brain's ability to encode statistical properties in multiple reference frames contributes to robust visual processing.