Abstract Value Encoding in Neural Populations But Not Single Neurons.

Area of Science:

• Neuroeconomics
• Computational Neuroscience
• Decision Neuroscience

Background:

• Understanding how the brain represents value for economic choices is crucial.
• Existing models struggle to reconcile abstract comparison with concrete value details.

Purpose of the Study:

• To investigate how the brain encodes the value of risky versus safe options.
• To determine if neural representations allow for comparison of dissimilar offers.

Main Methods:

• Examined neuronal responses in five value-encoding brain regions in male macaques.
• Analyzed neural codes for risky and safe options with identical subjective values.

Main Results:

• Found no overlap in neural codes for risky and safe options, even when subjectively equivalent.
• Observed distinct, semi-orthogonal encoding subspaces for risky and safe options.
• Demonstrated that these subspaces are linearly transformable, enabling comparison.

Conclusions:

• Neural representations preserve offer-specific details while allowing abstract comparison.
• This dual encoding supports flexible decision-making under changing circumstances.
• Highlights population coding principles for resolving neural representation problems.