Applying fuzzy petri nets for evaluating the impact of bedtime behaviors on sleep quality

Abstract

Sleep is an essential human activity, and medical studies have found that long-term poor sleep quality can contribute to the development of mental disorders and cardiovascular disease. The electroencephalogram (EEG) is a direct indicator of states of consciousness and brain function and can be used to detect sleep quality. This study develops a model for evaluating sleep quality through the fast Fourier transform and fuzzy Petri nets. The proposed EEG-based sleep quality detection model was found to provide an 82.4% increase in assessment accuracy when compared against other data mining methods. Moreover, this study explored the effect of various pre-sleep activities on subsequent sleep quality. Using mobile electronic devices for Internet browsing, watching videos, or playing games before sleeping was found that changes in brainwave power occur in the α, β, θ, and the overall frequency bands (observations of brain activity are often explained in terms of different frequency bands, alpha, beta, and theta bands are 8~13 Hz, 13~30 Hz, and 4~8 Hz, respectively), resulting in a moderate deterioration of sleep quality. Among these activities, watching videos had the most significant effect on sleep quality for both males and females. Furthermore, all three tested bedtime activities were found to have a more pronounced effect on the sleep quality of female subjects, while watching videos had a disproportionate effect on male subjects.