研究目的
To verify that the level of drowsiness produced by the POG-based drowsiness monitoring system is well related with two references: (1) the level of drowsiness obtained by analyzing polysomnographic signals; and (2) the performance of individuals in the accomplishment of Psychomotor Vigilance Tests (PVTs). A secondary goal is to determine the best threshold among the POG-based levels of drowsiness to alert individuals before they become dangerous.
研究成果
The POG-based drowsiness monitoring system effectively quantifies drowsiness levels, showing strong correlations with PSG-based measures and performance impairments. A threshold of 5 on the 0-10 scale is optimal for predicting lapses, with high sensitivity and specificity. The system is non-invasive, task-independent, and has potential for preventing accidents in various applications.
研究不足
The validation was performed in laboratory conditions; further evaluation in real-world settings (e.g., driving) is needed. The threshold for alerting may vary depending on the application and individual differences in drowsiness susceptibility.
1：Experimental Design and Method Selection:
The experiment involved 24 healthy volunteers performing Psychomotor Vigilance Tests (PVTs) under different sleep conditions over two days to assess drowsiness levels. The POG-based system was compared to polysomnography (PSG) and PVT performance as references.
2：Sample Selection and Data Sources:
Participants were selected based on criteria including age 18-35 years, no drug addiction, no sleep pathologies, not shift workers, and no jet lag in the preceding two weeks. Data included images of the right eye, PSG signals, and PVT data.
3：List of Experimental Equipment and Materials:
Equipment included a POG drowsiness monitoring system (prototype of Drowsimeter R100 from Phasya), actigraph (Actiwatch 2, Philips Respironics), and devices for PSG and PVT.
4：Experimental Procedures and Operational Workflow:
Participants underwent sleep deprivation and performed PVTs at specific times. Ocular parameters were extracted from eye images in 20-s windows, and drowsiness levels were computed automatically. PSG signals were analyzed for KDS scores, and PVT performance was measured by reaction times and lapses.
5：Data Analysis Methods:
Statistical analyses included repeated ANOVA to assess effects of sleep deprivation, Pearson's correlation coefficients to relate POG-based levels to references, and ROC curve analysis to determine the best drowsiness threshold.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
