Abstract

The aim of this research was to study if clustering of sleep characteristics of medical students at King George Medical University, Lucknow has any role in the effect of flaxseed on Total Sleep Time (TST). Methods: Polysomnographic and DLMO data of Ninety-six medical students who were randomly selected have been used in the study. Cluster analysis is a powerful multivariate tool for detecting structures in datasets. In the absence of known classification variable or the kind of relationship between observations, the method of cluster analysis, an unsupervised method is popularly used for knowing subgroups in data. Using agglomerative hierarchical clustering, we grouped data with complete linkage aggregation method on mutual similarity. The goal was to divide the students into groups with relatively homogeneous anthropometric and objective sleep features and study the role of consuming flaxseed for one month on students’ sleep time. As an intervention the students were given flaxseeds to eat daily continuously for 30 days. After this treatment, they were again assessed on their sleep characteristics. Results: Using complete linkage methodology on the pre-treatment sleep characteristics of 96 students, we obtained four clusters dominated by objective sleep characteristics whose centroids were significantly different. TST in clusters is bifurcated at 5 hrs sleep. Influence of flaxseeds on the total sleep time was obtained in first three clusters. Post treatment centroid of third cluster with about 5.5 hours TST was significantly higher than first two clusters. Conclusion: Mebsolt tetrad variables BMI, TST, sleep latency and DLMO might help to form subgroups of students to study the effect of flaxseed on TST.

Description