Abstract
Objective To develop a new fMRI network inference method, BrainNET, that utilizes an efficient machine learning algorithm in a specialized way to quantify contributions of various regions of interests (ROIs) in the brain to a specific ROI to estimate the network.
Methods BrainNET is based on Extremely Randomized Trees (ERT) to estimate network topology from fMRI data and modified to generate an adjacency matrix representing brain network topology, without reliance on arbitrary thresholds. Open source simulated fMRI data of fifty subjects in twenty-eight different simulations under various confounding conditions with known ground truth was used to validate the method. Performance was compared with Pearson correlation. The real-world performance was then evaluated in a publicly available Attention-deficit/hyperactivity disorder (ADHD) dataset including 135 Typically Developing Children (mean age: 12.00, males: 83), 75 ADHD Inattentive (mean age: 11.46, males: 56) and 93 ADHD Combined (mean age: 11.86, males: 77) subjects. Network topologies were inferred using BrainNET and Pearson correlation. Graph metrics were extracted to determine differences between ADHD groups. An extension to BrainNET was also developed (B-Corr) in which BrainNET adjacency matrix is combined with Pearson correlation output to remove false positives.
Results BrainNET demonstrated excellent performance across all simulations and varying confounders. It achieved significantly higher accuracy and sensitivity than Pearson correlation (p<0.05). In ADHD dataset, BrainNET was able to identify significant changes (p< 0.05) in graph metrics between groups. No significant changes in graph metrics between ADHD groups was identified using Pearson correlation. The B-Corr method provided similar results to BrainNET.