High-dimensional structure underlying individual differences in naturalistic visual experience.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Neural representations of sensory input vary significantly across individuals.
• The underlying architecture driving these individual differences in visual processing is not well understood.

Purpose of the Study:

• To investigate how unique visual experiences arise from identical sensory input.
• To explore the high-dimensional neural geometry underlying inter-individual variability in the visual cortex.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to record brain activity during naturalistic movie viewing.
• Spectral decomposition of fMRI responses was applied to analyze neural patterns across multiple dimensions.
• Intersubject correlation measures were used for comparison.

Main Results:

• Idiosyncratic neural patterns were found to persist across various orders of magnitude of latent dimensions.
• Distinct dimensional ranges within the neural geometry encoded qualitatively different aspects of individual visual processing.
• This multidimensional neural geometry predicted behavioral differences in memory recall and narrative description abstractness.

Conclusions:

• Subjective visual experience emerges from information integrated across an expansive, high-dimensional neural manifold.
• A geometric framework of neural activity provides a novel approach to understanding individual differences in perception.
• These findings challenge conventional measures of inter-individual variability and highlight the complexity of subjective visual worlds.