Pathak et al's new study proposes a novel structural organization for the brain, schematically represented above. Here, densely connected brain regions are arranged into sequential layers with the greatest connectivity between regions in adjacent layers. This allows for an interplay between distributed and sequential modes of information processing.
Scientists have discovered that common design principles underlie the connections within brains of varying complexity. This map of brain organization can help us understand the routes taken to process information, which allow organisms to respond to the world around them.
A recent study from the Complex Systems and Data Science (CSDS) research group at The Institute of Mathematical Sciences, Chennai has come up with a robust method to detect hierarchical organization in complex networks. Applying this metric to worm, macaque and human brain networks, the team discovered common patterns of brain organization. They found a signature of ‘modular hierarchy’, where the brain network is organized into individual well-connected groups, each of which is made up of layers with sequential connections. This common blueprint suggests that animal brains may process information in parallel, while also integrating outputs across them.
The brain can be seen as a network, known as a connectome, where the smallest unit (node) is a nerve cell or a brain region, and the connections between them are synapses or neural tracts between regions. “Studying the brain as a network is a mathematical tool to find regular patterns in its seemingly entangled wiring,” says Anand Pathak, the lead author of the study. The structure of this network can be described using two useful properties. Modularity is related to the grouping of nodes, where nodes within a group are more well-connected to each other than nodes in other groups. Hierarchy is another property where nodes or groups are arranged in a way that the output of one level serves as an input to the next. However, the measurement of this property in brain networks is often obscured by ‘shortcuts’ – connections between very different hierarchical levels, says Sitabhra Sinha, an author of the study who leads the CSDS research group.
To detect these structural properties, the team devised an index, which measures hierarchy as the extent of connectivity between adjacent layers in a network. They used a computer algorithm to search across various rearrangements of nodes and layers to find one that maximizes this index for a given network.
“We were surprised to discover that the networks did not just exhibit a hierarchical structure, but that these hierarchical levels were embedded within modules,” says Shakti Menon, an author of the study. “Modular structure has been well studied in the brain but the fact that it is intertwined with the hierarchical structure is a fresh new insight,” says Pathak. Another surprising aspect was that this template is shared between the worm Caenorhabditis elegans with about 300 nerve cells and macaque and human brains with billions of nerve cells, Pathak adds.
“A network that is modular in nature is optimized for distributed processing, while the flow is more sequential in a hierarchical network,” says Menon. “Despite their vast differences in complexity, all of these brain networks are organized to reap the benefits of two very different types of information processing mechanisms.” These insights support our current understanding and intuition about how the brain functions. From this study, we know that adjacent layers in a network hierarchy are, in fact, neighbouring regions in the brain. And their sequence of arrangement matches our current knowledge of how information flows in the brain.
This method would be useful for understanding the structure of other complex networks with similar properties, says Sinha. For example, it can help us tease apart the different trophic tiers in complex food webs and understand the structure of animal social networks, he adds.
Reference: Pathak, A., Menon, S. N., & Sinha, S. (2024). A hierarchy index for networks in the brain reveals a complex entangled organizational structure. Proceedings of the National Academy of Sciences USA, 121(27), e2314291121. https://doi.org/10.1073/pnas.2314291121
