Abstract Humans spend a significant portion of their time in the state of sleep, and therefore one's’sleep health’ is an important indicator of the overall health of an individual. Non‐invasive methods such as electroencephalography (EEG) are used to evaluate the ’sleep health’ as well as associated disorders such as nocturnal front lobe epilepsy, insomnia, and narcolepsy. A long‐duration and repetitive activity, known as a cyclic alternating pattern (CAP), is observed in the EEG waveforms which reflect the cortical electrical activity during non‐rapid eye movement (NREM) sleep. The CAP sequences involve various, continuing periods of phasic activation (phase‐A) and deactivation (phase‐B). The manual analysis of these signals performed by clinicians are prone to errors, and may lead to the wrong diagnosis. Hence, automated systems that can classify the two phases (viz. Phase A and Phase B accurately can eliminate any human involvement in the diagnosis. The pivotal aim of this study is to evaluate the usefulness of stopband energy minimized biorthogonal wavelet filter bank (BOWFB) based entropy features in the identification of CAP phases. We have employed entropy features obtained from six wavelet subbands of EEG signals to develop a machine learning (ML) based model using various supervised ML algorithms. The proposed model by us yielded an average classification accuracy of 74.40% with 10% hold‐out validation with the balanced dataset, and maximum accuracy of 87.83% with the unbalanced dataset using ensemble bagged tree classifier. The developed expert system can assist the medical practitioners to assess the person's cerebral activity and quality of sleep accurately.

Nowadays, the hectic work life of people has led to sleep deprivation. This may further result in sleep-related disorders and adverse physiological conditions. Therefore, sleep study has become an active research area. Sleep scoring is crucial for detecting sleep-related disorders like sleep apnea, insomnia, narcolepsy, periodic leg movement (PLM), and restless leg syndrome (RLS). Sleep is conventionally monitored in a sleep laboratory using polysomnography (PSG) which is the recording of various physiological signals. The traditional sleep stage scoring (SSG) done by professional sleep scorers is a tedious, strenuous, and time-consuming process as it is manual. Hence, developing a machine-learning model for automatic SSG is essential. In this study, we propose an automated SSG approach based on the biorthogonal wavelet filter bank's (BWFB) novel least squares (LS) design. We have utilized a huge Wisconsin sleep cohort (WSC) database in this study. The proposed study is a pioneering work on automatic sleep stage classification using the WSC database, which includes good sleepers and patients suffering from various sleep-related disorders, including apnea, insomnia, hypertension, diabetes, and asthma. To investigate the generalization of the proposed system, we evaluated the proposed model with the following publicly available databases: cyclic alternating pattern (CAP), sleep EDF, ISRUC, MIT-BIH, and the sleep apnea database from St. Vincent's University. This study uses only two unipolar EEG channels, namely O1-M2 and C3-M2, for the scoring. The Hjorth parameters (HP) are extracted from the wavelet subbands (SBS) that are obtained from the optimal BWFB. To classify sleep stages, the HP features are fed to several supervised machine learning classifiers. 12 different datasets have been created to develop a robust model. A total of 12 classification tasks (CT) have been conducted employing various classification algorithms. Our developed model achieved the best accuracy of 83.2% and Cohen's Kappa of 0.7345 to reliably distinguish five sleep stages, using an ensemble bagged tree classifier with 10-fold cross-validation using WSC data. We also observed that our system is either better or competitive with existing state-of-art systems when we tested with the above-mentioned five databases other than WSC. This method yielded promising results using only two EEG channels using a huge WSC database. Our approach is simple and hence, the developed model can be installed in home-based clinical systems and wearable devices for sleep scoring.

Abstract Congestive heart failure (CHF) is a cardiac disorder caused due to inefficient pumping of the heart, which leads to insufficient blood flow to the various parts of the body. The electrocardiogram (ECG) is widely used for the detection of heart diseases. However, it is prone to noise resulting in the detection of P, Q, R, S, and T waves ambiguous and erroneous. The heart rate variability (HRV) is considered to be a good indicator of various cardiac abnormalities. Hence, HRV is preferred. HRV can depict the magnitude of pumping of the heart in the RR interval signals accurately. This work proposes a method to automatically identify CHF using two‐band stopband energy (SBE) optimized orthogonal wavelet filter bank with HRV signals. In the proposed method, we have segmented the HRV data into lengths of 500 and 2000 samples. The HRV signals are decomposed into six sub‐bands, and the wavelet coefficients obtained are used for the extraction of fuzzy entropy (FE) and log energy (LE) features. The extracted features are utilized to classify HRV signals into control and CHF‐affected patients using support vector machine (SVM), bagged tree, complex tree, k‐nearest neighbour (KNN), and linear discriminant classifiers. The SVM performed better than other classifiers yielding the classification accuracy and maximum classification accuracy of 99.30% with (2000 samples) using cubic SVM (CSVM). The 10‐fold cross‐validation method is employed during classification to reduce the over‐fitting phenomenon (Sharma, Dhiman, & Acharya, 2021). It appears that the proposed optimal wavelet‐based automated system can identify CHF accurately using HRV signals. Hence, the model may be applied in clinical usage during an emergency employing a cloud‐based wireless system after testing the developed model with more data.

Abstract An early stage detection of Parkinson's disease (PD) is crucial for its appropriate treatment. The quality of life degrades with the advancement of the disease. In this paper, we propose a natural (time) domain technique for the diagnosis of PD. The proposed technique eliminates the need for transformation of the signal to other domains by extracting the feature of electroencephalography signals in the time domain. We hypothesize that two inter‐channel similarity features, correlation coefficients and linear predictive coefficients, are able to detect the PD signals automatically using support vector machines classifier with third degree polynomial kernel. A progressive feature addition analysis is employed using selected features obtained based on the feature ranking and principal component analysis techniques. The proposed approach is able to achieve a maximum accuracy of 99.1±0.1%. The presented computer‐aided diagnosis system can act as an assistive tool to confirm the finding of PD by the clinicians.