Statistical and Machine Learning Analysis of the Human Brain Functional Network in a Multi-Site Resting-State Functional MRI Database Framework

Authors

Oswaldo Artiles, Florida International University
Fahad Saeed, Ed., Florida International UniversityFollow

Date of this Version

2023

Document Type

Report

Abstract

The human brain has a complex network structure that is non-random and multiscale. It consists of subsystems coupled by a nonlinear dynamic, enabling it to produce complex responses to various external inputs and self-organize. To understand the physical structure and specific brain functions, it is essential to comprehend the connectivity of the hundreds of billions of neurons in the human brain. Functional connectivity (FC) in modern neuroscience is the statistical temporal dependencies between neuronal activation events occurring in spatially separated brain regions. Resting-state functional magnetic resonance imaging (rs-fMRI) is a non-invasive imaging technique widely used in neuroscience to understand the functional connectivity of the human brain. The studies presented in this dissertation were based on the models and methods from network neuroscience, which is an active area of research developed in the last three decades. These methods were used to model and analyze the functional human brain networks in a multi-site rs-fMRI data framework.

The contributions made in this dissertation to the study of the functional connectivity of the human brain network are:

1. The GPU-based Sparse Fast Fourier Transform (SFFT) of k-sparse signals;

2. The GPU-based breadth-first search algorithm;

3. The GPU-based betweenness centrality graph metric algorithm;

4. A comprehensive approach to solving the problem of confounding effects in the machine learning classification models of rs-fMRI multi-site data; and

5. A preliminary assessment of time-varying functional connectivity in a multisite data rs-fMRI framework. We hope that the neuroscience research community will use and improve these contributions to enhance the discovery of the functions and structure of the human brain. This will lead to a better understanding of the causes of brain disorders and the development of useful and effective biomarkers for their diagnosis.

Recommended Citation

Artiles, Oswaldo and Saeed,, Fahad Ed., "Statistical and Machine Learning Analysis of the Human Brain Functional Network in a Multi-Site Resting-State Functional MRI Database Framework" (2023). School of Computing and Information Sciences. 30.
https://digitalcommons.fiu.edu/cs_fac/30