Orthogonal neural representations support perceptual judgments of natural stimuli.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Visual Perception

Background:

• Natural vision involves separating relevant from irrelevant information.
• Previous studies often used artificial stimuli, not naturalistic images.
• Task-irrelevant background elements in natural scenes may impede object perception.

Purpose of the Study:

• To test if orthogonal neural representations support stable visual perception in natural scenes.
• To investigate the relationship between neural representations of object features and background features.
• To determine if irrelevant background features impair object position judgment.

Main Methods:

• Human psychophysics and monkey neurophysiology were employed.
• Parametrically variable naturalistic stimuli were used.
• Linear decoding models were applied to neural population responses in visual area V4.

Main Results:

• Neural representations of object position and background features in V4 were found to be orthogonal.
• Human observers' ability to judge object position was minimally affected by background features.
• Orthogonal representations were observed for various object and background features in V4.

Conclusions:

• Orthogonal neural representations in visual cortex can enable stable perception of object features amidst complex natural scenes.
• This neural coding strategy effectively segregates relevant from irrelevant visual information.
• Findings support a model where neural representations are optimized for robust visual processing in natural environments.