Although functional connectomics typically relies on resting-state fMRI, its analytical methods have been applied to task fMRI data in the investigation of broader involvements of brain regions even if inactive during a specific task. The purpose of this study is to assess the feasibility of inferring a true resting-state connectivity from task-fMRI data and to investigate the impact of connectomic-based analysis on behavioral trait studies. To this purpose, subjects underwent two visual fMRI tasks. The Blood-Oxygen-Level-Dependent (BOLD) time-series were processed to get both a "task" condition and a "pseudo-resting" condition applying different task regression setups to derive connectomes. Stimulus-classification experiments were conducted to compare "task" and "pseudo-resting" connectomes. Additionally, the influence of task regression was assessed through a classification experiment comparing children with Developmental Dyslexia (DD) and Typical Readers (TR). While task regression successfully removes task-related content from fMRI signals, stimulus information could still be inferred from connectomes, regardless of the preprocessing method used. Furthermore, a Support Vector Machine (SVM) experiment effectively discriminates between DD and TR in both "task" and "pseudo-resting" conditions. The study explored the impact of preprocessing in task fMRI experiments analyzed with connectomics. The ability to classify the stimuli in "pseudo-resting" conditions suggests that connectomes retain task-related signals even after task regression. Discriminative connections vary across tasks, affecting how classifiers differentiate between DD and TR. Despite these task-related differences, preprocessing had no effect on the inference of classification rules, indicating that key features are similarly evaluated in both tasks.
原文链接
分享
引用
