Investigation of the Effects of age and Gender Differences on Brain Activity Analysing Complexity Measures of Sleep EEG Signals by Machine Learning

Machine learning applications are used in the health sector, and as a result, many easy and applicable methods have been developed for analysis, interpretation, diagnosis and treatment. With the use of these applications in the sleep area, progress is expected in the diagnosis and treatment of sleep disorders, which is a major problem for millions of people. In order to achieve this goal, understanding the effects of age and gender factors that significantly affect brain activity during sleep, is important for all studies to be conducted. In this study, Lempel-Ziv complexity measures calculated from all-night sleep electroencephalogram signals were used in age and gender classification. Artificial Neural Networks (ANN) model has been used as the basic classifier in the studies. For ANN classifier, firstly, a network model was coded using MATLAB, and then a network model was designed using MATLAB nntool application for performance comparison. With the Classification Learner application, other classifier models were trained and their performances were examined, Decision Trees and Ensemble Classifiers showing the most successful results were included in the study. At the end of the study, age classification with 54.5% accuracy and general gender classification with 66.7% accuracy were achieved with the ANN model created using the MATLAB.