Variance-dependent neural activity in an unvoluntary averaging task

Area of Science:

• Neuroscience
• Cognitive Psychology
• Visual Perception

Background:

• Ensemble statistics, like mean estimation, allow for rapid gist perception of visual scenes.
• Averaging is crucial for estimating means and reducing variance, which can be external (stimulus-based) or internal (perceptual noise).
• The equivalent noise paradigm assumes averaging efficiency is constant regardless of variance source, a key assumption for measuring internal noise.

Purpose of the Study:

• To compare neural activity during visual averaging when variance originates from the stimulus versus internal estimation noise.
• To investigate the validity of the equivalent noise paradigm by examining neural correlates of different variance sources.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was employed to measure brain activity.
• Participants performed a visual averaging task involving Gabor orientations with varying levels of stimulus variance (high) and internal noise (low).

Main Results:

• Significantly greater activation was observed in the right superior frontal and left middle frontal gyri when stimulus variance was high compared to low variance conditions.
• These findings suggest distinct neural processing for external versus internal sources of variance during visual averaging.

Conclusions:

• The observed differences in neural activity challenge the core assumption of the equivalent noise paradigm.
• Future research using this paradigm must account for potential influences of differential neural activity on averaging efficiency, especially in high-variance scenarios.