Search

• A multi-domain feature extraction coupled with classification algorithm are used to detect the k-complexes. • Katz's algorithm, PSD and TQWT are employed to extract the EEG features. • The LS-SVM classifier is used to detect the k-complexes patterns in EEG. • This method provides better detection accuracy compared to other existing approaches. • The obtained results showed that the proposed method outperformed the others. Sleep scoring is one of the primary tasks for the classification of sleep stages using electroencephalogram (EEG) signals. It is one of the most important diagnostic methods in sleep research and must be carried out with a high degree of accuracy because any errors in the scoring in the patient's sleep EEG recordings can cause serious problems. The aim of this research is to develop a new automatic method for detecting the most important characteristics in sleep stage 2 such as k-complexes based on multi-domain features. In this study, each EEG signal is divided into a set of segments using a sliding window technique. Based on extensive experiments during the training phase, the size of the sliding window is set to 0.5 s (s). Then a set of statistical, fractal, frequency and non-linear features are extracted from each epoch based on the time domain, Katz's algorithm, power spectrum density (PSD) and tunable Q-factor wavelet transform (TQWT). As a result, a vector of twenty-two features is obtained to represent each EEG segment. In order to detect k-complexes, the extracted features were analysed for their ability to detect the k-complex waveforms. Based on the analysis of the features, twelve out of twenty-two features are selected and forwarded to a least square support vector machine (LS-SVM) classifier to identify k-complexes in EEG signals. A set of various classification techniques of K-means and extreme learning machine classifiers are used to compare the obtained results and to evaluate the performance of the proposed method.The experimental results showed that the proposed method, based on multi-domain features, achieved better recognition results than other methods and classifiers. An average accuracy, sensitivity and specificity of 97.7 %, 97 %, and 94.2 % were obtained, respectively, with the CZ-A1 channel according to the R&K standard. The experimental results with high classification performance demonstrated that the technique can help doctors optimize the diagnosis and treatment of sleep disorders. [ABSTRACT FROM AUTHOR]

Human brain functions and behaviors during the transition state between sleep and wakefulness are not similar to these at alert wakefulness state. The transition state, which is called sleep inertia, has many unpleasant and dangerous effects on many situations that require full attention and fast response, such as driving. Within 30 minutes after waking up from sleep, the driver's performance might be impaired due to the sleep inertia effects. Groups of drivers that may drive within a short period after waking up are: workers who travel early in the morning; secondary drivers of long distance bus who sleep in the bus before taking over the job from the primary drivers; night travelers; and long haul truck drivers who stop at the rest area to sleep for a while and continue driving. Previous research used subjective self-report measurement, eye tracker, and a driving simulator to analyze the driver's performance during sleep inertia state. The physiological measures of the drivers, such as their brain signals have also been studied. However, the brain signals which are recorded in Electroencephalography (EEG) are typically analyzed in perspective of the power spectrum. This study proposes a hybrid of EEG features, which are fractal dimension and power spectrum, supported by behavioral data which is the driver's reaction time. This study finds the features that significantly differentiate between normal and sleep inertia drivers based on the classification accuracy and p-value of the statistical ANOVA. This study compares the results with other features (power spectrum, variance, sample entropy), and between EEG channel. This study record the EEG from the Fz, T7, Cz, Pz, and O1 channels. This study uses subjective and behavioral measurements to support the results. The results show that the hybrid of fractal dimension estimated by Katz's algorithm at the O1 channel and delta power from the Fz channel, and alpha power from the O1 channel, have better classifications than the power spectrum alone. Furthermore, the reaction time recorded from the LED reaction time task shows a significant difference between drivers with sleep inertia and normal (alert) drivers.