Warren Howard J M, Ioachim Gabriela, Powers Jocelyn M, Stroman Patrick W
Centre for Neuroscience Studies, Queen's University, Kingston, Ontario, Canada.
Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Ontario, Canada.
J Pain Res. 2021 Feb 10;14:381-398. doi: 10.2147/JPR.S290795. eCollection 2021.
The purpose of this study was to investigate the utility of data-driven analyses of functional magnetic resonance imaging (fMRI) data, by means of structural equation modeling, for the investigation of pain processing in fibromyalgia (FM).
Datasets from two separate pain fMRI studies involving healthy controls (HC) and participants with FM were re-analyzed using both a conventional model-driven approach and a data-driven approach, and the results from these analyses were compared. The first dataset contained 15 women with FM and 15 women as healthy controls. The second dataset contained 15 women with FM and 11 women as healthy controls.
Consistent with previous studies, the model-driven analyses did not identify differences in pain processing between the HC and FM study groups in both datasets. On the other hand, the data-driven analyses identified significant group differences in both datasets.
Data-driven analyses can enhance our understanding of pain processing in healthy controls and in clinical populations by identifying activity associated with pain processing specific to the clinical groups that conventional model-driven analyses may miss.
本研究旨在通过结构方程模型，对功能磁共振成像（fMRI）数据进行数据驱动分析，以研究纤维肌痛（FM）患者的疼痛处理情况。
使用传统的模型驱动方法和数据驱动方法，对来自两项独立的疼痛fMRI研究的数据集进行重新分析，这两项研究涉及健康对照（HC）和FM患者，并比较这些分析的结果。第一个数据集包含15名FM女性患者和15名健康对照女性。第二个数据集包含15名FM女性患者和11名健康对照女性。
与先前的研究一致，模型驱动分析未在两个数据集中的HC和FM研究组之间发现疼痛处理方面的差异。另一方面，数据驱动分析在两个数据集中均发现了显著的组间差异。
数据驱动分析可以通过识别与临床组特定的疼痛处理相关的活动，增强我们对健康对照和临床人群中疼痛处理的理解，而传统的模型驱动分析可能会遗漏这些活动。
