Heterogeneous and higher-order cortical connectivity undergirds efficient, robust, and reliable neural codes.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Network Science

Background:

• Biological neural networks face a fundamental tradeoff between robustness and efficiency.
• Network architecture is hypothesized to play a key role in managing this tradeoff.
• Understanding how different network structures contribute to distinct functions is crucial.

Purpose of the Study:

• To investigate how the heterogeneous architecture of biological neural networks regulates the robustness-efficiency tradeoff.
• To develop a method for distinguishing neural subpopulations based on their connectivity patterns.
• To determine if different subpopulations optimize for efficiency versus robustness.

Main Methods:

• Developed a novel metric based on simplicial complex theory to quantify connectivity complexity.
• Validated the metric using multiple open-access connectome datasets.
• Employed a biologically detailed cortical model and electron microscopy data for analysis.

Main Results:

• Subpopulations with low simplicial complexity demonstrate efficient neural activity.
• Subpopulations with high simplicial complexity contribute to overall network reliability.
• The study confirmed the coexistence of both efficient and reliability-focused subpopulations within a single connectome.

Conclusions:

• Heterogeneous connectivity in neural networks allows for specialized subpopulations.
• Simplicial complexity serves as a useful metric for identifying functional roles of neural subpopulations.
• The findings provide insights into how the brain balances competing demands of efficiency and robustness.