| 摘要 |
This paper presents electroencephalographic (EEG)-based methods for assessing mental fatigue induced by
a prolonged vigilance task. After enrollment, 20 healthy male volunteers performed a mental fatigue-inducing
experiment. The experiment consisted of ten 9-min blocks of fatigue-inducing instrument monitoring, and after
each block volunteers completed subjective scales of fatigue, meanwhile the task performance of each block was
recorded. Over the course of the task, EEG readings were recorded and analyzed. Along with lower performance
and an increase in subjective levels of fatigue, there was a significant increase of ? energy in the frontal, central,
parietal, and occipital regions of the brain, and a decrease of ? energy in the frontal, inferior frontal, central,
posterior temporal, and occipital regions. The ? and ? energies were relatively steady, only increasing in the
occipital region. Significant increases of (?+??/?, ?/?, (?+??/(?+??, ?/?, and wavelet entropy were found in all
cerebral regions except the temporal region, where only ?/? was obviously changed. A high-dimensional vector
was constructed using these EEG indicators and cerebral regions, based on which the support vector machine
(SVM) was used to classify mental fatigue. The classifier for distinguishing between fatigue and non-fatigue was
91.1% accurate. This study demonstrates the feasibility of an EEG-based method for assessing and monitoring
mental fatigue during a vigilance task and provides theoretical basis for the future study. |