基于离散小波变换的意识障碍患者意识水平评估  

Assessment of Consciousness Level of Patients with Disorders of Consciousness Based on Discrete Wavelet Transform

作　　者：毛培名 黑玉光 杨姗姗 季虹 MAO Pei-ming;HEI Yu-guang;YANG Shan-shan;JI Hong(School of Computer Science,Xi'an Polytechnic University,Xi'an 710600,China;The Shaanxi Key Laboratory of Clothing Intelligence,School of Computer Science,Xi'an Polytechnic University,Xi'an 710600,China)

机构地区：[1]西安工程大学计算机科学学院,陕西西安710600 [2]西安工程大学计算机科学学院陕西省服装设计智能化重点实验室,陕西西安710600

出　　处：《计算机技术与发展》2024年第11期14-20,共7页Computer Technology and Development

基　　金：国家自然科学基金项目(62106189)

摘　　要：意识障碍(DOC)包括不同反应水平的神经状态,如无反应觉醒综合症(又称植物人状态,UWS/VS)和最小意识状态(MCS)。准确的对意识障碍患者进行意识水平评估,有助于医生为意识障碍患者提供合适的康复治疗方案,以最大程度恢复或改善患者的意识水平。脑电图能够实时记录反映意识障碍患者意识水平的生理电活动,是实现意识水平评估的重要工具。基于此,该文利用离散小波变换对意识障碍患者的静息态脑电图数据进行处理,以获取包含不同脑电频带信息的多尺度信号。然后,从这些多尺度信号中提取了不同生理脑电图频带的关键特征——功率、谱熵和相干度。最后,采用融合特征结合支持向量机(SVM)分类器实现意识障碍患者的意识水平评估。实验结果显示,该方法获得了91.37%的分类准确率。与已有的意识水平评估方法相比较,该方法在意识障碍患者分类识别方面表现出更高的准确率。Disorders of consciousness(DOC) encompass various neurological states with different levels of responsiveness,such as the unresponsive wakefulness syndrome(UWS/VS) and the minimally conscious state(MCS).Accurate assessment of consciousness level in DOC patients is crucial for physicians to devise appropriate rehabilitation strategies aimed at maximizing consciousness recovery or improvement.Electroencephalography(EEG) serves as a vital tool in real-time monitoring of physiological electrical activity,reflecting the consciousness levels of DOC patients,thus facilitating consciousness assessment.In this regard,we utilized discrete wavelet transform to process resting-state EEG data from DOC patients,extracting multiscale signals containing information across different EEG frequency bands.Subsequently,key features – power,spectral entropy,and coherence – were extracted from these multiscale signals corresponding to different physiological EEG frequency bands.Finally,consciousness levels of DOC patients were evaluated using a support vector machine(SVM) classifier with feature fusion.Experimental results demonstrated a classification accuracy of 91.37% using the proposed method.Compared to existing consciousness assessment approaches,the proposed method exhibited higher accuracy in classifying and identifying consciousness levels of DOC patients.

关 键 词：意识障碍 意识水平评估 静息态脑电图 离散小波变换 支持向量机 融合特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]