An Investigation on Eye Tracking Indicators to Detect Alertness on Inspection Jobs with Sleep Limits and Gender Differences

Abstract

Eye tracking (ET) has been widely used to detect ocular changes that are also correlated with fatigue. However, whether there is an ET indicator capable of detecting fatigue closely related to vigilance decrement is unknown. This study aims to determine which ET indicators are sensitive in detecting changes in vigilance due to sleep restriction (4 and 8 hours) in different gender. Fifty-six participants (28 men, 21.32 ± 1.16 years and 28 women, 21.5 ± 1.2 years) were involved in this study. Using an eye tracker device, participants were asked to simulate inspecting defects (dirt, scratches, missing parts, and poor assembly) at 50 flashlights. In addition, vigilance assessment was carried out with the Psychomotor Vigilance Task (PVT) and the Karolinska Sleepiness Scale (KSS). The results of ANOVA showed that sleep restriction affected the mean RT (p-value = 0.001), minor lapse (p-value = 0.036), mean 1/RT (p-value = 0.001), saccadic velocity (p-value = 0.001), and KSS (p-value = <0.001). Gender influenced the mean 1/RT (p-value = 0.016), pupil diameter (p-value = 0.022), and saccadic velocity (p-value = 0.030). The Pearson correlation test shows a weak relationship between the ET and PVT indicators. However, saccadic velocity has the potential to be able to detect changes in vigilance because it is influenced by sleep restriction. Further research is still needed to confirm these results because the correlation between saccade velocity and PVT indicators in this study still needs to show a stronger relationship to become a vigilance indicator.