It's the network stupid! // EIT Digital

It's the network stupid!

Researchers applied control theory, normally used in engineered systems, to predict which regions could exert the most influence. Credit: University of California Riverside

Small worlds always attracted my imagination.  This is a (relatively novel) branch of mathematics, and systems theory, that studies the relevance of networks in a variety of systems. It turns out that their study explains many properties of the world around us, and inside us. It explains the resiliency of connected systems and how beyond a critical point they crumble, how a flock or a swarm seems to take up a life of its own, and how intelligent behaviour emerges from seemingly simple "things".

Hence you can imagine my interest in reading that researchers at University of California, Riverside, have looked into the brain connectivity to discover how this leads to cognitive control.

They have used imaging techniques to identify the main connection path within a human brain creating wiring diagrams among the various parts of the brain and have applied network rules to see how cognitive control can occur. It has been know since several decades that the frontal part of our brain, the frontal cortex, is at the core of our cognitive capability. The study shows, from a mechanistic point of view why this is the case.

The work has been published on Nature Communications and it is a marriage of neuroscience with the growing field of network science. Network science is not, as one might assume, a telecommunication derived discipline. Telecommunications networks, in a way, are too simple to make emergent properties occur. Network science investigates social networks where volume and variability create very complex environment. This is something that will eventually happen, according to several scientists (and although I don't qualify as scientist I have the same feeling), once networks will explode in two direction: volume, by incorporating "things" (Internet of things) and total flexibility (SDN and NFV as basic architecture supporting autonomous systems/nodes).

Interestingly, the study at Riverside shows that (contrary to what one would assume) only a decentralised region, like the frontal cortex, can dynamically control the flow of thoughts and goal directed behaviour. A region "inside the brain" would not be able to do that. The reason is that thoughts and cognition is not the result of processing (where a central processing would make good engineering sense) but it is the result of networks.

According to one of the researchers, Danielle Basset, the brain "working" resembles a flock of birds. The flock has a direction and change direction pushed by infinitesimal stimuli from each individual bird, without any single bird playing a role of orchestrator. A tiny variation in a bird flight might get amplified to the point that it affects the flock or might be dampened and disappear. The likelihood of one or the other happening is also a function of where a bird is within the network. Something very similar occur in the brain. Different regions because of their connectivity play a different role in steering, making our thoughts emerge.

Distant regions, like the frontal cortex, are more effective in diverting your thoughts significantly, whilst central regions are less likely to do that. These latter can push to small transitions. Again, we find small world concepts (weak and strong links, where most of impact is actually connected to weak links, not to strong ones).

Interesting stuff! And I bet we will be discovering and applying these discovery in many areas, from smart cities to autonomous driving cars 

Author - Roberto Saracco

© 2010-2018 EIT Digital IVZW. All rights reserved. Legal notice. Privacy Policy.