Brain Connectivity Studies on Structure‐Function Relationships: A Short Survey with an Emphasis on Machine Learning

Abstract

This short survey reviews the recent literature on the relationship between the brain structure and its functional dynamics. Imaging techniques such as diffusion tensor imaging (DTI) make it possible to reconstruct axonal fiber tracks and describe the structural connectivity (SC) between brain regions. By measuring fluctuations in neuronal activity, functional magnetic resonance imaging (fMRI) provides insights into the dynamics within this structural network. One key for a better understanding of brain mechanisms is to investigate how these fast dynamics emerge on a relatively stable structural backbone. So far, computational simulations and methods from graph theory have been mainly used for modeling this relationship. Machine learning techniques have already been established in neuroimaging for identifying functionally independent brain networks and classifying pathological brain states. This survey focuses on methods from machine learning, which contribute to our understanding of functional interactions between brain regions and their relation to the underlying anatomical substrate.

The virtual aging brain: a model-driven explanation for cognitive decline in older subjects

“…There have been various attempts in the literature trying to explain the complex relationship between SC and FC (Fukushima et al, 2018;Suárez et al, 2020;Zimmermann et al, 2016Zimmermann et al, , 2019, but they all relied on data-driven and statistical methods, which limited the interaction between the two modalities to percolation (Saggio et al, 2016;Wein et al, 2021). Straathof et al (Straathof et al, 2019) pointed out that the estimation of the SC-FC relationship may be characterized by a large unexplained variability possibly due the limited information of the regional averaged FC.…”

Generative Models of Brain Dynamics -- A review

“…One common formulation is to build the dynamical graph models of the cortex based on the anatomical, functional, or effective connectivity as described in Table 1. For a more comprehensive review of such networks, refer to [121].…”

“…The machine learning techniques are now routinely used for classification and regression of brain states (see Wein et al [121] for a review). However, they have much more potential than black-box, data-intensive classifiers.…”

Forecasting Brain Activity Based on Models of Spatio-Temporal Brain Dynamics: A Comparison of Graph Neural Network Architectures

“…For learning the predictions of the BOLD signal, samples of input and output sequences were generated from the timeseries data in X [78]. This was achieved by selecting windows of length T p to obtain input sequences of neural activity states [x (1) , .…”

“…These approaches have already contributed numerous insights into the relationship between functional and structural brain networks, like those that explain how functional coherency patterns emerge in biophysically inspired models constrained by anatomical connections [36,53], or how indirect structural connections contribute to the inference of FC strength [47,9]. Therefore the vast majority of studies focuses on inferring overall FC patterns from their SC, although static coherency based measures of FC might have limitations in their ability to capture the rich nature of dynamic brain activity [78]. To the con-trary, STGNNs are able to directly predict the measured BOLD dynamics, and their interactions between brain regions, without relying on the indirect representation of functional dynamics based on coherency.…”