The feasibility of a Drowsiness Monitoring Device (DMD) to detect EEG indices of drowsiness in real- time, was demonstrated during Phase I with an analytical model correctly classifying 97.1 percent of sleep episodes and 94.3 percent of awake epochs in 20 sleep-deprived subjects. The model employs discriminant function analysis (DFA) to characterize and classify one-sec epochs, validated against a combination of visual scoring by polysomnographers and/or a behavioral measure. Algorithms to detect artifacts in real-time (EMG, 60-Hz and gross body/eye movements) were developed. This classification accuracy represents a significant advancement over previously reported models and confirms the feasibility of distinguishing EEG characteristics of sleep and waking on a second-by- second basis. Phase II will implement a three-level DFA classification system to further refine the model, adding sub-class states for vigilance and drowsiness/sleep to improve system accuracy. The multi- dimensional time-series DFA analyses will correlate EEG parameters with behavioral measures of driving performance to provide quantitative predictions of performance decrements associate with sleep onset. The model will be validated using a population with demographics consistent with the target market for the DMD (e.g., truck drivers) during Phase II. In addition, the system will be evaluated with the introduction of commonly used legal drugs (caffeine, nicotine, and cold medications) to determine the robustness of the model.
The DMD provides three levels of user safety. When sleep onset is approaching, the DMD will initiate a verbal warning alarm that must be turned off by the user. Alternatively, the DMD can provide verbal feedback to ensure the user maintains high levels of alertness during activities that require sustained vigilance. The user can also select the option for the DMD to recommend the optimal time to take a short nap, monitor the length of the nap and awaken the user at the appropriate time. Currently, more that 10% of the U.S. workforce or an estimated 20 million people are engaged in night sift work. The transportation industry, including airline, railroad, marine and highway transportation companies, is the nation's third largest employer of shift workers. Long haul truck drivers, in particular, are vulnerable to sleepiness because they drive through the night, in most cases unaccompanied, and generally sleep less than 6 hours per day at irregular intervals. In addition, an estimated 6 million Americans suffer from chronic sleep disorders which make them vulnerable to fatigue in the workplace.
| Johnson, Robin R; Popovic, Djordje P; Olmstead, Richard E et al. (2011) Drowsiness/alertness algorithm development and validation using synchronized EEG and cognitive performance to individualize a generalized model. Biol Psychol 87:241-50 |
