Denoising strategies of functional connectivity MRI data in lesional and non-lesional brain diseases

Abstract Functional connectivity magnetic resonance imaging (fcMRI) is a widely utilized tool for analyzing functional connectivity (FC) in both healthy and diseased brains. However, patients with brain disorders are particularly susceptible to head movement during scanning, which can introduce substantial noise and compromise data quality. Therefore, identifying optimal denoising strategies is essential to ensure reliable and accurate downstream data analysis for both lesional and non-lesional brain conditions. In this study, we analyzed data from four cohorts: healthy subjects, patients with brain lesions (glioma, meningioma), and patients with a non-lesional encephalopathic condition. Our goal was to evaluate various denoising strategies using quality control (QC) metrics to identify the most effective approach for minimizing noise while preserving the integrity of the blood oxygen level-dependent (BOLD) signal, tailored to each disease type. The effectiveness of denoising strategies varied based on the data quality and whether the data were derived from lesional or non-lesional diseases. At comparable levels of head motion, combinations involving independent component analysis-based automatic removal of motion artifacts (ICA-AROMA) denoising strategies were most effective for data from a non-lesional encephalopathic condition, while combinations including anatomical component correction (CC) yielded the best results for data from lesional conditions. Here, we present the first comparison of denoising pipelines for patients with lesional and non-lesional brain diseases. A key finding was that, at comparable levels of head motion, the optimal denoising strategy varies depending on the nature of the brain disease.