Liu Xue, Niu Peining, He Jinchen, Du Guowei, Xu Yan, Liu Tao, Yang Zhaoxu, Liu Shaowei, Chen Yun, Chen Jianhuai
Department of Andrology Jiangsu Province Hospital of Chinese Medicine Affiliated Hospital of Nanjing University of Chinese Medicine Nanjing China; Chengdu University of Traditional Chinese Medicine Chengdu China.
Department of Andrology Siyang Traditional Chinese Medicine Hospital Suqian China.
Neuroscience. 2025 Feb 16;567:219-226. doi: 10.1016/j.neuroscience.2025.01.015. Epub 2025 Jan 10.
Psychogenic erectile dysfunction (pED) is often accompanied by abnormal brain activities. This study aimed to develop an automaticclassifier to distinguish pED from healthy controls (HCs) by identified brain-basedcharacteristics. Resting-state functional magnetic resonance imaging data were acquired from 45 pED patients and 43 HCs. Regional homogeneity (ReHo) and functional connectivity (FC) values were calculated and compared between groups. Moreover, based on altered ReHo and FC values, support vector machine (SVM) classifier, incorporating recursive feature elimination (RFE), an SVM-RFE diagnostic model was established using leave-one-out cross-validation. Patients demonstrated reduced ReHo values in the left middle temporal gyrus (had decreased FC values with the left medial superior frontal gyrus and cuneus), orbital part of inferior frontal gyrus (had decreased FC values within the same region), triangular part of inferior frontal gyrus, anterior cingulate gyrus (had decreased FC values with the left inferior temporal gyrus, anterior cingulate gyrus, cuneus and right supplementary motor area) and middle frontal gyrus. The right calcarine fissure displayed increased ReHo values. The diagnostic model demonstrated excellent performance, achieving an accuracy rate of 90.80%. This study identified altered regional activity and FC in specific brain regions of pED patients, which might be related to the development of pED. The application of machine learning confirmed the distinctive characteristics of these functional changes in the brain. The high accuracy of our diagnostic model suggested a promising direction for developing objective diagnostic tools for psychological disorders.
心因性勃起功能障碍（pED）常伴有大脑活动异常。本研究旨在通过识别基于大脑的特征来开发一种自动分类器，以区分pED患者与健康对照者（HCs）。从45例pED患者和43例HCs中获取静息态功能磁共振成像数据。计算并比较两组之间的局部一致性（ReHo）和功能连接（FC）值。此外，基于ReHo和FC值的改变，采用支持向量机（SVM）分类器并结合递归特征消除（RFE），使用留一法交叉验证建立了SVM-RFE诊断模型。患者在左侧颞中回（与左侧额上回内侧和楔叶的FC值降低）、额下回眶部（同一区域内的FC值降低）、额下回三角部、前扣带回（与左侧颞下回、前扣带回、楔叶和右侧辅助运动区的FC值降低）和额中回的ReHo值降低。右侧距状裂的ReHo值升高。该诊断模型表现出优异的性能，准确率达到90.80%。本研究确定了pED患者特定脑区的区域活动和FC改变，这可能与pED的发生有关。机器学习的应用证实了大脑中这些功能变化的独特特征。我们诊断模型的高准确率为开发心理障碍的客观诊断工具提供了一个有前景的方向。
