基于头动与眼动的脑疲劳检测方法研究

Research on Mental Fatigue Detection Method Based on Head Movement and Eye Movement

目的为实现准确、可靠地检测脑疲劳,研究基于头动与眼动特征判断受试者脑疲劳程度的方法。方法采集16名男性受试者在36h睡眠剥夺(SD)实验中的头动与眼动信号,进行去噪预处理,并基于主成分分析法(PCA)融合头动与眼动特征,综合判断受试者脑疲劳程度,并对由脑电信号判断的脑疲劳程度、双重任务作业绩效、警戒作业任务模拟测试(PVT)反应时间、主观瞌睡度分值进行比对。结果基于头动与眼动特征能很好地判断受试者脑疲劳程度,且与通过脑电信号判断的脑疲劳程度相关性为0.771±0.030;与双重任务作业绩效变化的相关性为0.665±0.024;与PVT反应时间的相关性为0.812±0.011;与主观瞌睡度分值的相关性为0.682±0.023,且均显著相关(P<0.05)。结论基于PCA融合头动与眼动特征能有效、准确检测受试者脑疲劳程度,与传统检测方法相比具有很好的一致性。

Objective In order to detect the mental fatigue accurately and reliably,a method to estimate the degree of mental fatigue based on the characteristics of head movement and eye movement was studied.Methods The head movement and eye movement signals of 16 male subjects in 36 hours sleep deprivation(SD)were collected and pre-denoised.The head movement and eye movement characteristics were integrated based on the principal component analysis(PCA)method.The subjects’mental fatigue was estimated,and the changes of mental fatigue evaluated by EEG signals,dual task performance,vigilance task simulation test(PVT)response time,and subjective sleepiness scores were recorded for comparison.Results Based on the characteristics of head movements and eye movements,the degree of mental fatigue of the subjects was well estimated,and the correlation with the degree of mental fatigue estimated by the EEG signal was 0.768±0.033;the correlation with the change in performance of dual task work was 0.663±0.020;the correlation with PVT response time was 0.814±0.009;the correlation with subjective sleepiness score was 0.762±0.023,and all were significantly correlated(P<0.05).Conclusion PCA based fusion of head movement and eye movement features can effectively and accurately detect the degree of mental fatigue of subjects,which has a good consistency with the traditional detection methods.