论文部分内容阅读
目的研究睡眠-清醒状态下大脑血氧变化特性,探讨基于功能近红外光谱深层次信息识别两状态的可行性。方法采用便携式功能近红外设备(波长849 nm、757 nm)测量了6名年轻(20~26岁)男性志愿者静止躺卧姿势下睡眠-清醒状态大脑前额位置6通道的功能近红外谱(f NIRs)数据,计算得到反映前额叶血氧水平变化的氧合血红蛋白(HbO_2)均值、心动信号和Burg功率谱等共计36个生理特征,并用支持向量机(support vector machine,SVM)分类器建立了清醒与睡眠两个状态的识别模型。结果在清醒与睡眠过渡阶段大脑前额叶HbO_2均值水平呈下降趋势,心动信号频率在睡眠由浅入深过程中也呈下降趋势;睡眠时功率谱中呼吸波与心动强度均比清醒时有所下降(符合正常睡眠呼吸变缓、心动减弱规律);睡眠-清醒状态的平均分类识别准确率可达90%。结论基于功能近红外光谱信息检测实现人体静止躺卧姿势下睡眠与清醒状态的自动化识别具有技术可行性,对在轨航天员空间作息评估与规划具有实际应用价值。
Objective To investigate the changes of cerebral blood oxygenation during sleep-wakefulness and to explore the feasibility of identifying two states based on deep-level information of functional near-infrared spectroscopy. Methods The function of near infrared (F (subscript f)) channel of six young (20-26 years old) male volunteers in sleep-wake state at forehead lying position in resting-lying position was measured by using portable functional near- infrared device (wavelength 849 nm, 757 nm) NIRs) were used to calculate the mean HbO2, heartbeat signal and Burg power spectrum which reflect the changes of prefrontal lobe blood oxygen level. A total of 36 physiological traits were calculated and analyzed by the support vector machine (SVM) Awake and sleep two state recognition model. Results The level of HbO_2 in the prefrontal cortex of the brain during the period of awake and sleep showed a decreasing trend. The frequency of heartbeat signal also showed a downward trend during the process of going from deep to deep. During sleep, the respiratory wave and heartbeat intensity in the power spectrum were lower than those in the awake In line with normal sleep breathing slowed, reduced heart motility); sleep-awake state average classification accuracy of 90%. Conclusion It is feasible to realize the automatic recognition of sleep and wakefulness under static lying position based on the detection of functional near-infrared spectral information, and has practical application value for the space astronaut in life assessment and planning.