Optimal coding predicts attentional modulation of activity in neural systems

Area of Science:

• Computational Neuroscience
• Cognitive Neuroscience
• Information Theory

Background:

• Neuronal activity changes with attentional state, suggesting the neural code adapts.
• Selective attention influences how the brain processes sensory information.

Purpose of the Study:

• To propose an information-theoretic model for attentional modulation of neural activity.
• To explain how the brain optimizes sensory encoding based on feature relevance.

Main Methods:

• Computer simulations of a feedforward coding system.
• Development of a model learning a covert attention mechanism.
• Analysis of coding strategy reorganization based on attentional signals.

Main Results:

• The model demonstrates attentional modulation without multiplicative interactions or gating.
• Simulated activity modulation matches experimental observations in selective attention.
• The model predicts stimulus-dependent attentional modulation functions.

Conclusions:

• Attentional modulation arises from optimal coding principles, not specialized attention mechanisms.
• This information-theoretic approach provides a general framework for understanding attention across modalities.
• The model offers insights into how computational resources are reallocated during attention.