Enhancing Functional Connectivity Analysis in Task-Based fMRI Using the BOLD-Filter Method: Greater Network and Activation Voxel Sensitivities

Abstract

Task-based functional MRI (tb-fMRI) has gained prominence for investigating brain connectivity by engaging specific functional networks during cognitive or behavioral tasks. Compared to resting-state fMRI (rs-fMRI), tb-fMRI provides greater specificity and interpretability, making it a valuable tool for examining task-relevant networks and individual differences in brain function. In this study, we evaluated the utility of the BOLD-filter--a method originally developed to extract reliable BOLD (blood oxygenation level-dependent) components from rs-fMRI--by applying it to tb-fMRI data as a preprocessing step for functional connectivity (FC) analysis. The goal was to enhance the sensitivity and specificity of detecting task-induced functional activity. Compared to the conventional preprocessing method, the BOLD-filter substantially improved the isolation of task-evoked BOLD signals. It identified over eleven times more activation voxels at a high statistical threshold and more than twice as many at a lower threshold. Moreover, FC networks derived from BOLD-filtered signals revealed clearer task-related patterns, including gender-specific differences in brain regions linked to everyday behaviors. These patterns were not detectable using standard preprocessing approaches. Our findings demonstrate that the BOLD-filter enhances the robustness and interpretability of FC analysis in tb-fMRI. By effectively isolating meaningful functional networks, this approach offers significant advantages over conventional preprocessing methods. The BOLD-filter holds promise for advancing both basic neuroscience research and clinical applications by enabling more precise characterization of task-induced brain activity.