Seeing white: Qualia in the context of decoding population codes

Area of Science:

• Neuroscience
• Computational Neuroscience
• Sensory Perception

Background:

• The nervous system integrates multiple simultaneous inputs (e.g., wavelength, disparity) to generate novel percepts.
• Existing neural population decoding models struggle with multiple inputs, either extracting single values or failing to synthesize new representations.
• This limitation highlights a broader challenge in interpreting population codes.

Purpose of the Study:

• To propose a new framework for understanding neural population decoding.
• To demonstrate that decoding involves transformation into an abstract representational space, not just physical parameter extraction.
• To present a computational model illustrating this transformation.

Main Methods:

• Developed a four-layer neural network model.
• The network transforms input wavelength data.
• The transformation targets a high-level hue-saturation color space representation.

Main Results:

• The proposed network successfully transforms wavelength inputs into a distinct color space.
• This demonstrates a mechanism for generating novel percepts from combined inputs.
• The model shows that decoding can lead to abstract representations beyond simple physical parameters.

Conclusions:

• Neural population decoding is better understood as a transformation to abstract representational spaces.
• This approach can explain the generation of complex percepts from multiple sensory inputs.
• The proposed model offers a new perspective on interpreting neural population codes.