Efficient coding theory of dynamic attentional modulation.

Area of Science:

• Computational neuroscience
• Systems neuroscience

Background:

• Sensory neuron activity is influenced by external stimuli and internal feedback from higher brain areas.
• Attention is a key internal signal hypothesized to optimize sensory representations for relevant information at minimal neural cost.
• This hypothesis lacked a normative computational framework.

Purpose of the Study:

• To develop a normative computational model of dynamic population coding in the visual cortex.
• To investigate how attention-like modulation emerges from principles of efficient coding and probabilistic inference.
• To explore the implications for understanding cortical phenomena.

Main Methods:

• Developed a computational model based on normative principles of probabilistic inference and efficient coding.
• Simulated dynamic population coding in the visual cortex.
• Analyzed the emergent properties of the sensory code under changing perceptual demands.

Main Results:

• An attention-like modulation emerged from the model by continuously adapting the sensory code.
• This modulation significantly reduced neural activity.
• Task-specific inference accuracy was maintained despite reduced neural activity.

Conclusions:

• The model demonstrates how attention can be understood as an emergent property of efficient coding and probabilistic inference.
• This framework unifies diverse cortical phenomena, including gain modulation, tuning modulation, and response variability.
• These findings suggest a common underlying principle for sensory processing in dynamic environments.