基于功能近红外光谱的多生理脑力疲劳检测

发布时间：2018-10-23 08:49

【摘要】：脑力疲劳会引起人机系统绩效下降甚至引起安全事故,因此实时检测疲劳状态具有重要意义。虽然关于脑力疲劳检测的研究较多,但仍未有统一生理标准。由于疲劳的复杂性,多生理检测法已经成为一种趋势,但是会增大设备复杂度。功能近红外光谱能够通过测量人大脑皮层的血氧活动而间接反映脑认知功能,近红外信号中的心动和呼吸信号属于生理活动的敏感信息,但是常被作为干扰去除,因此造成了信息丢失。为增强近红外的生理信息含量并建立多生理疲劳检测模型,从近红外信号中提取出心动和呼吸作为新的敏感特征,并结合均值斜率等常规特征构建基于支持向量机的脑力疲劳检测模型。研究采用60 min 2-back任务诱导疲劳状态,利用近红外测量了15名被试包括前额(PFC)共计10个通道的脑皮层近红外信号。研究结果证实了提取出的心动和呼吸特征对疲劳敏感,且增大了疲劳识别的准确性(84%→90%)。因此,所建立的模型能够有效地检测脑力疲劳并且降低了多生理脑力疲劳检测设备的复杂度。
[Abstract]:Mental fatigue can cause deterioration of man-machine system performance and even cause safety accidents, so it is very important to detect fatigue state in real time. Although there are many researches on mental fatigue detection, there is still no unified physiological standard. Due to the complexity of fatigue, multi-physiological detection has become a trend, but it will increase the complexity of equipment. Functional near infrared spectroscopy (FNIR) can indirectly reflect the cognitive function of brain by measuring the blood oxygen activity of human cerebral cortex. The cardiac and respiratory signals in NIR signal are sensitive information of physiological activities, but they are often removed as interference. As a result, information is lost. In order to enhance the physiological information content of NIR and establish a multi-physiological fatigue detection model, cardiac and respiratory signals were extracted from NIR signals as a new sensitive feature. A mental fatigue detection model based on support vector machine (SVM) was constructed based on the conventional features such as mean slope. The 60 min 2-back task induced fatigue state was used to measure the cortical near infrared signals of 15 subjects, including 10 channels of prefrontal (PFC). The results show that the extracted cardiac and respiratory characteristics are sensitive to fatigue and increase the accuracy of fatigue identification (84% or 90%). Therefore, the established model can effectively detect mental fatigue and reduce the complexity of multiple physiological mental fatigue detection equipment.
【作者单位】：中国航天员科研训练中心;
【基金】：国家自然科学基金(81671861) 中国航天医学工程预先研究项目(YJGF151204) 中国航天员科研训练中心人因国家重点实验室自主课题(SYFD150051805)项目资助
【分类号】：R318;TN219

【参考文献】