| 基于深度学习与多模态医学影像融合识别阈下抑郁患者 |
| |
| 引用本文： | 尹小龙,李德民,图娅,单保慈. 基于深度学习与多模态医学影像融合识别阈下抑郁患者[J]. 中国医学影像技术, 2020, 36(8): 1158-1162 |
| |
| 作者姓名： | 尹小龙 李德民 图娅 单保慈 |
| |
| 作者单位： | 郑州大学物理学院, 河南 郑州 450001;北京中医药大学针灸推拿学院, 北京 100029;北京师范大学认知神经科学与学习国家重点实验室, 北京 100875;中国科学院高能物理研究所 北京市射线成像技术与装备工程技术研究中心, 北京 100049;中国科学院大学核科学与技术学院, 北京 100049 |
| |
| 基金项目： | 国家自然科学基金面上项目（81671770）。 |
| |
| 摘 要： | 目的 基于深度学习（DL）卷积神经网络（CNN）算法，利用医学影像数据实现识别阈下抑郁（StD）患者。方法 对56例StD患者和70名正常人采集MRI和fMRI数据，分别输入所构建的CNN，利用网络融合技术对2种不同模态数据进行综合分析，得到分类结果；最后调整网络结构与模型参数，实现分类效果最优化。结果 单独MRI数据模型分类精度为73.02%，单独fMRI数据模型分类精度为65.08%；2种模态结合，最终分类精度升至78.57%。结论 利用DL可识别StD患者与正常人；采用多种模态输入法可提高分类准确度。
|
| 关 键 词： | 抑郁 深度学习 多模态显像 磁共振成像 |
| 收稿时间： | 2019-08-21 |
| 修稿时间： | 2020-03-17 |
Identification of subthreshold depression based on deep learning and multimodal medical image fusion |
| |
| YIN Xiaolong,LI Demin,TU Y,SHAN Baoci. Identification of subthreshold depression based on deep learning and multimodal medical image fusion[J]. Chinese Journal of Medical Imaging Technology, 2020, 36(8): 1158-1162 |
| |
| Authors: | YIN Xiaolong LI Demin TU Y SHAN Baoci |
| |
| Affiliation: | School of Physics and Engineering, Zhengzhou University, Zhengzhou 450001, China;School of Acupuncture-Moxibustion and Tuina, Beijing University of Chinese Medicine, Beijing 100029, China;State Key Laboratory of Cognitive Neuroscience and Learning-BNU, Beijing 100875, China; Beijing Engineering Research Center of Radiographic Techniques and Equipment, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China;School of Nuclear Science and Technology, University of Chinese Academy of Sciences, Beijing 100049, China |
| |
| Abstract: | Objective To explore the value of convolutional neural network (CNN) algorithm based on deep learning (DL) for identification of subliminal depression (StD) patients using medical image data. Methods MRI and fMRI data of 56 StD patients (StD group) and 70 normal controls(NC group) were collected and input into the constructed CNN, respectively. Then the network fusion technology was used to comprehensively analyze the two different modalities to obtain the classification result. Finally, the network fusion technology was used to integrate two different modal data and optimize the classification effect. Results The identification accuracy of the structural image data alone was 73.02%, of the functional image data alone was 65.08%. With combination of the two modes, the final classification accuracy raised to 78.57%. Conclusion DL can classify patients with StD and normal subjects. Multiple modal input methods can improve classification accuracy. |
| |
| Keywords: | depression deep learning multimodal imaging magnetic resonance imaging |
|
| 点击此处可从《中国医学影像技术》浏览原始摘要信息 |
|
点击此处可从《中国医学影像技术》下载全文 |
|
