Guan Jianwu, Li Hai, Yang Qiansu, Lv Yanwei, Zhang Lei, Wang Yi, Li Shijun
Department of Radiology, First Medical Center, Chinese PLA General Hospital, Beijing, 100871, China.
Center for MRI Research, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, 100871, China.
Commun Biol. 2025 Mar 21;8(1):473. doi: 10.1038/s42003-025-07928-w.
Head motion during magnetic resonance imaging (MRI) examinations of patients with autism spectrum disorder (ASD) can influence the identification of brain differences as well as early diagnosis and precise MRI-based interventions for ASD. This study aims to address head motion issues in resting-state functional MRI (rs-fMRI) data by comparing various correction methods. Specifically, we evaluate the independent component analysis-based automatic removal of motion artifacts (ICA-AROMA) against traditional preprocessing pipelines, including head motion realignment parameters and global signal regression (GSR). Our dataset consisted of 306 participants, including 148 individuals with ASD and 158 participants with typical development (TD). We find that ICA-AROMA, particularly when combined with GSR and physiological noise correction, outperformed other strategies in differentiating ASD from TD participants based on functional connectivity (FC) analyses. The correlation of quality control with functional connectivity (QC-FC) is statistically significant in proportion and distance after applying each denoising pipeline. The mean FC between groups is significant for Yeo's 17-Network in each denoising strategy. ICA-AROMA head motion correction outperformed other strategies, revealing more significant FC networks and distinct brain regions linked to the posterior cingulate cortex and postcentral gyrus. This suggests ICA-AROMA enhances fMRI preprocessing, aiding ASD diagnosis and biomarker development.
在对自闭症谱系障碍（ASD）患者进行磁共振成像（MRI）检查时，头部运动可能会影响大脑差异的识别，以及ASD的早期诊断和基于MRI的精确干预。本研究旨在通过比较各种校正方法来解决静息态功能MRI（rs-fMRI）数据中的头部运动问题。具体而言，我们将基于独立成分分析的运动伪影自动去除方法（ICA-AROMA）与传统的预处理流程进行评估比较，传统流程包括头部运动重新对齐参数和全局信号回归（GSR）。我们的数据集包含306名参与者，其中包括148名ASD个体和158名发育正常（TD）的参与者。我们发现，ICA-AROMA，特别是与GSR和生理噪声校正相结合时，在基于功能连接（FC）分析区分ASD和TD参与者方面优于其他策略。在应用每种去噪流程后，质量控制与功能连接性（QC-FC）的相关性在比例和距离上具有统计学意义。在每种去噪策略中，Yeo的17网络中两组之间的平均FC具有显著性。ICA-AROMA头部运动校正优于其他策略，揭示了与后扣带回皮质和中央后回相关的更显著的FC网络和不同的脑区。这表明ICA-AROMA增强了fMRI预处理，有助于ASD诊断和生物标志物开发。
