Increase in Mutual Information During Interaction with the Environment Contributes to Perception

Area of Science:

• Neuroscience
• Information Theory
• Computational Neuroscience

Background:

• Perception and motor interaction rely on neuronal activity, specifically the binary states (spikes or no spikes) of neurons.
• Changes in probability laws governing these binary states, particularly entropy reduction, are linked to processing environmental stimuli.
• Existing research suggests mutual information increases with sensory processing across cortical areas and networks.

Purpose of the Study:

• To propose that increased mutual information between sensory inputs and neural circuit connectivity underlies perception and motor control.
• To investigate the role of mutual information in coupling action and perception.
• To explore how processing sensory information without prior knowledge impacts Shannon information and surprise.

Main Methods:

• Analysis of probability distributions of neuronal binary states.
• Quantification of entropy reduction and mutual information in neural circuits.
• Examination of information processing across hierarchical cortical areas and networks.

Main Results:

• Perception and motor interaction are associated with reduced entropy in neuronal binary states.
• Increased mutual information, reflecting sensory input characteristics and neural connectivity, is proposed as a key mechanism.
• Mutual information increases during sensory processing, particularly between lower and higher cortical areas or different networks.

Conclusions:

• Robust increases in mutual information are crucial for perception and successful motor interactions with the physical environment.
• Mutual information links sensory stimuli, neural processing, and motor responses, explaining the action-perception coupling.
• Processing novel sensory inputs increases Shannon information and surprise, strengthening the neural model of the environment.