利用脑功能网络与临床特征对慢性意识障碍进行预后预测
神经元与认知
2018-09-10 v3
摘要
意识障碍是不同疾病或损伤的混合异质体。尽管已有一些指标和模型被提出用于预后预测，但任何单一方法单独使用时都存在较高的误判风险。本研究旨在开发一种多域预后模型，将静息态功能磁共振成像与三项临床特征相结合，在单受试者水平上预测一年后的转归。该模型在两个医学中心的三个数据集上，以约90%的准确率区分了后续恢复意识与未恢复意识的患者。它还能够识别不同预测因子（包括脑功能与临床特征）的预后重要性。据我们所知，这是首个基于静息态功能磁共振成像与临床特征的慢性意识障碍多域预后模型的实现。因此我们认为，这一新型预后模型准确、稳健且可解释。
引用
@article{arxiv.1801.03268,
title = {Prognostication of chronic disorders of consciousness using brain functional networks and clinical characteristics},
author = {Ming Song and Yi Yang and Jianghong He and Zhengyi Yang and Shan Yu and Qiuyou Xie and Xiaoyu Xia and Yuanyuan Dang and Qiang Zhang and Xinhuai Wu and Yue Cui and Bing Hou and Ronghao Yu and Ruxiang Xu and Tianzi Jiang},
journal= {arXiv preprint arXiv:1801.03268},
year = {2018}
}
备注
Although some prognostic indicators and models have been proposed for disorders of consciousness, each single method when used alone carries risks of false prediction. Song et al. report that a model combining resting state functional MRI with clinical characteristics provided accurate, robust, and interpretable prognostications. 52 pages, 1 table, 7 figures