A Proposal for Sleep Scoring Analysis Designed by Computer Assisted using Physiological Signals

Sleep is utterly regarded as compulsory componentfor a person’s prosperity and is an exceedingly important elementfor wellbeing of a healthy person. It is a condition in which anindividual is physically and mentally at rest. The conception ofsleep is considered extremely peculiar and is a topic of discussionand researchers all over the world has been attracted by thisconcept. Sleep analysis and its stages is analyzed to be useful insleep research and sleep medicine area. By properly analyzingthe sleep scoring system and its different stages has provenhelpful for diagnosing sleep disorders. As it’s seen, sleep stageclassification by manual process is a hectic procedure as it takessufficient time for sleep experts to perform data analysis. Besides,mistakes and irregularities in between classification of same datacan be recurrent. Therefore, the use of automatic scoring systemin order to support reliable classification is highly in greater use.The scheduled work provides an insight to use the automaticscheme which is based on real time EMG signals and Artificialneural network. EMG is an electro neurological diagnostic toolwhich evaluates and records the electrical activity generated bymuscle cells. The sleep scoring analysis can be applied byrecording Electroencephalogram (EEG), Electromyogram(EMG), and Electrooculogram (EOG) based on epoch and thismethod is termed as PSG test or polysomnography test. Theepoch measured has length segments for a period of 30 seconds.The standard database of EMG records was gathered fromvarious hospitals in sleep laboratory which gives the differentstages of sleep. These are Waking, Non-REM1 (stage-1), NonREM2 (stage-2), Non-REM3 (stage-3), REM. The collection ofdata was done for the period of 30 second known as epoch, forseven hours. The dataset obtained from the biological signal wasmanaged so that necessary data is to be extracted fromdegenerated signal utilized for the purpose of study. As a matterof fact, it is known electrical signals are distributed throughoutthe body and is needed to be removed. These unwanted signalsare termed as artifacts and they are removed with the help offilters. In this proposed work, the signal is filtered by making useof low-pass filter called Butterworth. The withdrawncharacteristics were instructed and categorized by utilizingArtificial Neural Network (ANN). ANN, on the other hand ishighly complicated network and utilizing same in the field ofbiomedical when contracted with electrical signals, acquiredfrom human body is itself a novel. The precision obtained by thehelp of the procedure was discovered to be satisfactory and hencethe process is very useful in clinics of sleep, especially helpful forneuro-scientists for discovering the disturbance in sleep.