光学学报, 2018, 38 (3): 0317001, 网络出版: 2018-03-20
基于功能性近红外光谱成像的警觉度检测 下载: 823次
Vigilance Detection Based on Functional Near-Infrared Spectroscopy
光谱学 功能性近红外光谱成像 警觉度 氧合血红蛋白 脱氧血红蛋白 spectroscopy functional near-infrared spectroscopy vigilance oxygenated hemoglobin deoxygenated hemoglobin
摘要
采用功能性近红外光谱成像(fNIRS)研究了警觉度的变化规律。通过10通道fNIRS技术采集了12名受试者前额部位的氧合血红蛋白(HbO)信号、脱氧血红蛋白(Hb)信号, 并记录下实验中受试者的行为学数据。结果表明:可以根据主观量表评分和行为学数据将受试者的警觉度标注为3个水平, 前30 min为高警觉度水平, 中间30 min为一般警觉度水平, 后30 min为低警觉度水平;此外, 前额左侧的第2通道以及右侧的第9、7、8、6通道对警觉度变化的敏感度较高;在支持向量机模型下, 12名受试者的警觉度三分类正确率为76.9%。本研究验证了fNIRS在警觉度检测上的可行性, 并指出了对警觉度变化敏感脑区的位置, 为警觉度的实时监测提供了新的思路。
Abstract
The change regulation of vigilance based on functional near-infrared spectroscopy (fNIRS) is studied. During the experiment, the hemodynamics signals of 12 participants including oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (Hb) are recorded through 10 channels fNIRS technology, and the behavior data of the participants are simultaneously recorded. The results indicate that vigilance of the participants can be labeled as three levels: the first 30 min is high vigilance level, the middle 30 min is general vigilance level, and the last 30 min is low vigilance level. Additionally, the results indicate that channel 2 on the left side and channel 9, channel 8, channel 7 and channel 6 on the right side of prefrontal have better sensitivity to vigilance change. Under support vector machine recognition model, the average accuracy over the 12 participants of three-level-classification results is 76.9%. The experimental results confirm the feasibility of fNIRS in the evaluation of vigilance, and point out the specific locations of the sensitive brain region, which provides a new way for real-time vigilance monitoring.
曹勇, 焦学军, 姜劲, 傅嘉豪, 潘津津. 基于功能性近红外光谱成像的警觉度检测[J]. 光学学报, 2018, 38(3): 0317001. Cao Yong, Jiao Xunjun, Jiang Jin, Fu Jiahao, Pan Jinjin. Vigilance Detection Based on Functional Near-Infrared Spectroscopy[J]. Acta Optica Sinica, 2018, 38(3): 0317001.