Representation of visual uncertainty through neural gain variability.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Visual Perception

Background:

• Perception inherently involves uncertainty, requiring neural circuits to represent information reliability.
• The precise mechanisms by which neural circuits encode uncertainty remain a subject of ongoing research and debate.

Purpose of the Study:

• To propose and test a model where neural response strength encodes stimulus features and cross-neuron gain variability encodes uncertainty.
• To investigate how neural circuits in the visual cortex represent and process stimulus uncertainty.

Main Methods:

• Studied spiking activity of neurons in macaque V1 and V2.
• Presented stimuli with manipulated uncertainty and analyzed neural responses.
• Examined the relationship between neural gain variability and stimulus uncertainty.

Main Results:

• Average neural response strength correlates with stimulus features.
• Cross-neuron response gain variability is tuned to stimulus uncertainty.
• This tuning is feature-specific and largely independent of the uncertainty's source.
• Observed that this neural behavior arises from known gain-control mechanisms.

Conclusions:

• Neural gain variability serves as a code for stimulus uncertainty in the visual cortex.
• Downstream circuits can decode both stimulus features and their uncertainty from population activity.
• The findings provide a framework for understanding how the brain handles uncertainty in sensory processing.