Holistic Atlases of Functional Networks and Interactions Reveal Reciprocal Organizational Architecture of Cortical Function

Abstract

For decades, it has been largely unknown to what extent multiple functional networks spatially overlap/interact with each other and jointly realize the total cortical function. Here, by developing novel sparse representation of whole-brain fMRI signals and by using the recently publicly released large-scale Human Connectome Project high-quality fMRI data, we show that a number of reproducible and robust functional networks, including both task-evoked and resting state networks, are simultaneously distributed in distant neuroanatomic areas and substantially spatially overlapping with each other, thus forming an initial collection of holistic atlases of functional networks and interactions (HAFNIs). More interestingly, the HAFNIs revealed two distinct patterns of highly overlapped regions and highly specialized regions and exhibited that these two patterns of areas are reciprocally localized, revealing a novel organizational principle of cortical function.

A novel framework for groupwise registration of fMRI images based on common functional networks

“…Inspired by the observation that multiple functional brain networks can be commonly identified across individuals via ICA methods [15] or dictionary learning algorithms [16], [17], common networks can be utilized to establish correspondence between fMRI images. In this paper, after preprocessing, functional ICNs are decomposed from each fMRI image and a set of z-score maps representing networks’ 3D pattern are obtained by FSL MELODIC ICA software (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/melodic).…”

Embedded sparse representation of fMRI data via group-wise dictionary optimization

“…Z is the coefficient matrix. An important characteristic of using sparse representation for fMRI signals is that the coefficient weight matrix (Z) naturally reveals the spatial patterns among those reconstructed brain networks [2,3]. Here, we adopted stochastic coordinate coding (SCC) [6] to solve this ordinary sparse learning problem.…”

“…Recent studies [3] suggest that some neuroscientifically meaningful dictionary atoms can always be achieved when the dictionary size lies within a reasonably wide range. In other words, some consistent dictionary atoms can be found across different subjects, even if each person's data is subjected to sparse learning separately.…”

“…Recently, there has been increasing interest in applying sparse representation methods to analyze functional magnetic resonance images of the brain [2][3]. Unlike other applications which pursue a compact or abstract representation of the original data, sparse learning of fMRI signals has additional justifications from a neuroscience point of view: though the exact mechanism is largely unknown, the human brain is widely considered to include a collection of specialized functional networks that flexibly interact when different brain functions are performed [4].…”

Multi-task sparse screening for predicting future clinical scores using longitudinal cortical thickness measures

“…In addition, over-fitting becomes one of the biggest concerns for building prediction models in brain imaging research. Recent developments from dictionary learning [14, 12, 13, 26] can offer valuable insights for above challenges. However, most existing works on dictionary learning focus on the prediction of a target outcome at a single time point [24, 22] or some region-of-interest [23, 25].…”