The connectivity of the brain: multi-level quantitative analysis

Area of Science:

• Computational Neuroscience
• Neuroanatomy
• Mathematical Modeling

Background:

• Understanding brain connectivity and neuronal organization is crucial for neuroscience.
• Previous models often simplified the complex interplay between neuronal topology and physical packing.

Purpose of the Study:

• To develop a mathematical framework for calculating brain connectivity volumes based on topology and physical packing.
• To compare model predictions with human neuroanatomical data across different brain regions and scales.
• To investigate interspecies differences in brain organization.

Main Methods:

• Developed a mathematical formalism to model connectivity volumes considering four topologies (full, random, nearest-neighbor, modular).
• Incorporated three physical packing strategies: interior, sheeted, and exterior.
• Utilized extensive human neuroanatomical data to derive consistent parameters for the whole brain, cerebral cortex, and cerebellar cortex.

Main Results:

• Exterior packing is more efficient than interior packing for spheres.
• Fully and randomly connected topologies are inefficient for the human brain.
• Modular topologies with exterior packing best fit macro-structural human brain data.
• Laminarization and columnarization (sheeted packing) are efficient at the mesostructural level.
• Interior models are most efficient within sheets at the microstructural level.
• Interspecies comparison suggests evolution minimizes interneuron distance in grey matter.
• Random topology may be feasible for mouse cortex, unlike human brain.

Conclusions:

• Human brain organization at different scales employs distinct topologies and packing strategies.
• Modular and exterior packing are favored at the macro-scale, while sheeted and interior packing dominate at meso- and micro-scales.
• Interspecies variations in brain topology, such as the feasibility of random connectivity in mice, highlight evolutionary adaptations.