Supervised approach based sleep disorder detection using non - Linear dynamic features (NLDF) of EEG

Abstract

Depending on its intensity, sleep disorders can affect a person's ability to function mentally, emotionally, and physically. These are medical abnormalities of the subject's sleep structure. Most prevalent ones are bruxism, sleep problems, depression, and narcolepsy. A greater chance of acquiring sleep issues in the elderly includes sleeplessness, irregular leg movements, problems with fast eye movement behaviour and breathing abnormalities. Therefore, an early stage therapy that might save a patient's life depends on a precise diagnosis and categorization. The much more sensitive as well as significant bio-signal is electroencephalographic (EEG) signal. It has the capacity to record sleep-sensitive brain activity. We used an available EEG database which had recordings divided into different types of sleep disturbances as well as a healthy control group. Popular sensor's EEG brain function has been examined. Ultimately, using patterns taken from EEG data, a categorization AI model was created. Extracted characteristics worked well as a biomarker for identifying sleep problems when combined with an AI classifier.

Keywords

• EEG
• Sleep disorder
• Bruxism
• Fractal dimension
• Lyapunov exponent