ABSTRACT
Sleep medicine defines the diagnosis and treatment of sleep-related disorders, the most common clinical examination being polysomnography (PSG), the simultaneous recording of multiple physiological signals during a night of sleep. The analysis of these signals can generate evidence to improve diagnosis and monitoring of sleep disorders. Big data approaches to sleep medicine are gaining increasing interest for identifying phenotypes, tailoring treatment, and predicting outcomes, although the large volume and variety of data pose a number of challenges. In this chapter, we present some strategies for the automatic analysis of biomedical signals recorded during PSG, with particular attention to the electroencephalogram, heart rate variability, and cardiorespiratory signals. We focus on techniques for addressing challenges such as data harmonization, missing data, and noise, and we present some examples of applications of big data approaches in sleep medicine. Throughout the chapter, we provide references to publically available PSG data sets and tools for PSG signal analysis.
