The spatiotemporal neural dynamics underlying perceived similarity for real-world objects

Area of Science:

• Cognitive Neuroscience
• Neuroimaging
• Visual Perception

Background:

• Perceived similarity influences object categorization and perception.
• Understanding neural mechanisms of similarity perception is crucial.
• Object dimensions often co-occur, necessitating analysis of unique dimensional roles.

Purpose of the Study:

• Investigate neural representations of perceived object similarity.
• Examine the roles of shape, function, color, and background in similarity perception.
• Relate behavioral similarity judgments to brain activity using fMRI and MEG.

Main Methods:

• Behavioral assessment of object similarity across dimensions (shape, function, color, background).
• Functional Magnetic Resonance Imaging (fMRI) and Magnetoencephalography (MEG) for neural activity.
• Representational Similarity Analysis (RSA) to link behavior and brain data.

Main Results:

• Object dimensions are linked to distinct, rapidly emerging (within 200ms) representations in the visual cortex.
• Color representations preceded shape representations in both processing hierarchy and time.
• Overlapping and distributed representations were observed across multiple dimensions.
• Shape representations were not fully explained by super-category or deep neural network models.

Conclusions:

• Neural representations of perceived object similarity are dimension-specific yet distributed.
• Visual cortex processes object dimensions like color and shape with distinct temporal dynamics.
• The study provides a comprehensive characterization of the relationship between object dimensions, perceived similarity, and neural activity.