Validation of a Light EEG-Based Measure for Real-Time Stress Monitoring during Realistic Driving

Abstract

Driver's stress affects decision-making and the probability of risk occurrence, and it is therefore a key factor in road safety. This suggests the need for continuous stress monitoring. This work aims at validating a stress neurophysiological measure-a Neurometric-for out-of-the-lab use obtained from lightweight EEG relying on two wet sensors, in real-time, and without calibration. The Neurometric was tested during a multitasking experiment and validated with a realistic driving simulator. Twenty subjects participated in the experiment, and the resulting stress Neurometric was compared with the Random Forest (RF) model, calibrated by using EEG features and both intra-subject and cross-task approaches. The Neurometric was also compared with a measure based on skin conductance level (SCL), representing one of the physiological parameters investigated in the literature mostly correlated with stress variations. We found that during both multitasking and realistic driving experiments, the Neurometric was able to discriminate between low and high levels of stress with an average Area Under Curve (AUC) value higher than 0.9. Furthermore, the stress Neurometric showed higher AUC and stability than both the SCL measure and the RF calibrated with a cross-task approach. In conclusion, the Neurometric proposed in this work proved to be suitable for out-of-the-lab monitoring of stress levels.

Keywords: EEG; driving; random forest; stress; wet EEG sensors.

Figure 1

Multitasking application screen. In clockwise order: the auditory monitoring, the phone number entry task, the visual monitoring and the mental arithmetic task.

Figure 2

Car simulator. The SERVER and CLIENT applications have been highlighted.

Figure 3

Boxplot representation of the performance in terms of percentage (left) and radar chart of NASA-TLX subscales (right). In the radar chart, the lines represent the median of each subscale and the shaded areas represent the maximum value in the population. The low-stress condition is represented in green, and high-stress in red. The asterisks show that p < 0.05 resulting from the Wilcoxon signed-rank test.

Figure 4

Boxplot representation of Area Under Curve (AUC) for seven different decimations (1 to 60 by 10 s step) of the multitasking experiment. The Random Forest (RF-intra) is represented in red, the Neurometric in blue, and the Skin Conductance Level (SCL) in orange. The points represent the outliers, and the asterisk shows that p < 0.05 resulting from the Dunn’s post-hoc tests.

Figure 5

Stress during the multitasking experiment. The time series were obtained from the three different models at three different decimations (10, 30 and 60 s). The line represents the average over the population of recorded stress value, the shadow represents the standard deviation. In grey the expected value of stress level (low- and high-stress conditions) is represented, the Random Forest (RF-intra) is represented in red, the Neurometric in blue, and the Skin Conductance Level (SCL) in orange.

Figure 6

Boxplot representation of the number of collisions (left) and radar chart of NASA-TLX subscales (right). In the radar chart the lines represent the median of each subscale and the shaded areas the maximum value in the population. The low-stress condition is represented in green, and high-stress in red. The asterisks show the p < 0.05 resulting from the Wilcoxon signed-rank test.

Figure 7

Boxplot representation of Area Under Curve (AUC) for seven different decimations (1 to 60 by 10 s step) of driving experiment. The Random Forest calibrated cross-task (RF-cross) is represented in green, the Random Forest calibrated intra-subject (RF-intra) in red, the Neurometric in blue, and the Skin Conductance Level (SCL) in orange. The points represent the outliers and the asterisk shows that p < 0.05 (the double-asterisk means p < 0.01) resulting from the Dunn’s post-hoc tests.

Figure 8

Stress during the driving experiment. The time series were obtained from the four different models at three different decimations (10, 30 and 60 s). The line represents the average over the population of recorded stress value, and the shadow represents the standard deviation. The expected value of stress level (low and high-stress condition) is represented in grey, the Random Forest calibrated cross-task (RF-cross) is represented in green, the Random Forest calibrated intra-subject (RF-intra) in red, the Neurometric in blue, and the Skin Conductance Level (SCL) in orange.