The Use of Principal Component Analysis for Reduction in Sleep Quality and Quantity Data in Female Professional Soccer

Abstract

The main aim of the present study was to uncover multivariate relationships between sleep quantity and quality using principal component analysis (PCA) in professional female soccer players. A second aim was to examine the extent to which objective sleep quantity and quality variables can discriminate between perceived sleep. Ten objective sleep variables from the multisensory sleep-tracker were analyzed. PCA was conducted on the sleep variables, and meaningful principal components (PCs) were identified (eigenvalue > 2). Two sleep PCs were identified, representing the ‘quantity of sleep’ (quantity PC: eigenvalue = 4.1 and variance explained = 45.1%) and the ‘quality of sleep’ (quality PC: eigenvalue = 2.4 and variance explained = 24.1%). Cluster analysis grouped the players’ sleep into three types: long and efficient, short and efficient, and long and inefficient; however, no association was found between the perceived sleep and the sleep clusters. In conclusion, a combination of both quantity and quality sleep metrics is recommended for sleep monitoring of professional female soccer players. Players should undergo a training process to improve self-assessment of sleep quality recorded from a subjective questionnaire, contrasting the perceived information with the sleep quality recorded objectively during a defined period in order to optimize the validity of their perceptions. The aim is to optimize the validity of their perceptions of sleep quality.

Keywords

• multivariate analysis
• sleep
• smart ring
• soccer