Leibnitz, K. | Search bioRxiv

bioRxiv2022

Packetization improves communication efficiency in brain networks with rapid and cost-effective propagation strategies

Computational studies in network neuroscience build models of communication dynamics in the con-nectome that help us understand the structure-function relationships of the brain. In these models, the dynamics of cortical signal transmission in brain networks are approximated with simple propagation strategies such as random walks and shortest path routing. Furthermore, the signal transmission dynamics in brain networks are associated with the switching architectures of engineered communication systems (e.g., message switching and packet switching). However, it has been unclear how propagation strategies and switching architectures are related in models of brain network communication. Here, we investigate the effects of the difference between packet switching and message switching (i.e., whether signals are packetized or not) on the transmission efficiency of the propagation strategies when simulating signal propagation in a macaque brain network. The results show that packetization decreases the efficiency of the random walk strategy and does not change the efficiency of the shortest path strategy, but increases the efficiency of more plausible strategies for brain networks that balance between communication speed and information cost. This finding suggests an advantage of packet-switched communication in the connectome and provides new insights into modeling the communication dynamics in brain networks.

neuroscience↗

bioRxiv2017

Classification and characterisation of brain network changes in chronic back pain: a multicenter study.

Chronic pain is a common and often disabling condition, and is thought to involve a combination of peripheral and central neurobiological factors. However, the extent and nature of changes in the brain is poorly understood. Here, we investigated brain network architecture using resting state fMRI data collected from chronic back pain patients in UK and Japan (41 patients, 56 controls). Using a machine learning approach (support vector machine), we found that brain network patterns reliably classified chronic pain patients in a third, independent open data set with an accuracy of 63%, whilst 68% was attained in cross validation of all data. We then developed a deep learning classifier using a conditional variational autoencoder, which also yield yielded 63% generalisation and 68% cross-validation accuracy. Given the existence of reliable network changes, we next studied the graph topology of the network, and found consistent evidence of hub disruption based on clustering and betweenness centrality of brain nodes in pain patients. To examine this in more detail, we developed a multislice modularity algorithm to identify a consensus pattern of modular reorganisation of brain nodes across the entire data set. This revealed evidence of significant changes in the modular identity of several brain regions, most notably including broad regions of bilateral sensorimotor cortex, subregions of which also contributed to classifier performance. These results provide evidence of consistent and characteristic brain network changes in chronic pain, and highlight extensive reorganisaton of the network architecture of sensorimotor cortex.

neuroscience↗