Efficient coding in the economics of human brain connectomics

Area of Science:

• Systems neuroscience
• Computational neuroscience
• Neuroimaging

Background:

• Models of brain function emphasize efficient information transfer within structural brain networks.
• Evidence for efficient communication, particularly in hierarchical networks with hubs, is limited.
• Efficient coding theory suggests maximal information transmission with minimal metabolic cost.

Purpose of the Study:

• To theorize and test how structural brain connectivity supports efficient coding.
• To introduce and apply a novel metric, compression efficiency, quantifying the fidelity-transmission trade-off.
• To investigate the developmental trajectory and behavioral relevance of compression efficiency.

Main Methods:

• Developed a theory of minimum transmission rates for expected fidelity using random walk dynamics.
• Introduced and calculated compression efficiency in structural brain networks.
• Analyzed diffusion-weighted imaging data from 1,042 youth (ages 8-23) and cerebral blood flow for metabolic expenditure.

Main Results:

• Structural brain networks exhibit compression efficiency trade-offs aligning with theoretical predictions.
• Compression efficiency prioritizes fidelity during development and is influenced by metabolic resources and myelination.
• This efficiency explains hierarchical organization benefits, links input fidelity to brain area expansion, and reveals hubs use lossy compression for integration.

Conclusions:

• Structural connectivity supports efficient coding through compression efficiency, a metric that refines understanding beyond traditional network efficiency.
• Compression efficiency predicts cognitive performance across executive function, memory, reasoning, and social cognition.
• Findings highlight the role of random walk dynamics in constrained communication within macroscale brain networks.