摘要
实时功能磁共振成像神经反馈(real time functional Magnetic Resonance Imaging neurofeedback, rtfMRI-NF)技术是治疗失眠障碍的新型手段,但是由于个体的差异,不同患者的治疗效果差异显著,利用脑影像数据对治疗效果进行预测,对提升神经反馈训练临床应用的个体适应性具有重要意义。在32名失眠患者进行rtfMRI-NF的实验基础上,提取了患者治疗前以左侧杏仁核为种子点的基于脑网络组模板的功能连接矩阵作为特征,构建机器学习预测模型,实现了利用支持向量机对患者rtfMRI-NF训练效果进行预测。实验结果表明,单个失眠患者是否受益于rtfMRI-NF情绪调节方法的预测准确率为87.5%,该方法能够初步为失眠患者选择个体化治疗方案时提供一定的参考依据。
Real time functional magnetic resonance imaging neurofeedback(rtfMRI-NF)is a new method for treating insomnia disorder.However,due to individual differences,there are significant differences in the therapeutic effect among patients.The prediction of therapeutic effect by using brain imaging data is of great significance for improving individual adaptability of clinical application of neurofeedback training.Based on the experiment of rtfMRI-NF in 32 insomnia patients,the func-tional connection matrix based on brainnetome template with the left amygdala as the region of inter-est before treatment was extracted as the feature and the machine learning prediction model was con-structed.Besides,the training effect of rtfMRI-NF in patients was predicted by support vector ma-chine.The experimental results show that the prediction accuracy of whether a single insomnia pa-tient benefits from rfMRI-NF emotion regulation method is 87.5%,which could provide a certain reference for insomnia patients to choose personalized treatment plan.
作者
李恺
张驰
童莉
闫镔
LI Kai;ZHANG Chi;TONG Li;YAN Bin(Information Engineering University,Zhengzhou 450001,China;Unit 61580,Beijing 100000,China)
出处
《信息工程大学学报》
2023年第6期699-704,共6页
Journal of Information Engineering University
基金
国家自然科学基金资助项目(82071884)。
关键词
失眠
神经反馈
功能连接
机器学习
insomnia
neurofeedback
functional connection
machine learning