BACKGROUND: Essential tremor (ET) and the tremor in Parkinson's disease (PD) are the two most common pathological tremor with a certain overlap in the clinical presentation. OBJECTIVE: The main purpose of this work is to use an artificial neural netw

BACKGROUND AND OBJECTIVE: Although careful clinical examination and medical history are the most important steps towards a diagnostic separation between different tremors, the electro-physiological analysis of the tremor using accelerometry and electromyography (EMG) of the affected limbs are promising tools. METHODS: A soft-decision wavelet-based decomposition technique is applied with 8 decomposition stages to estimate the power spectral density of accelerometer and surface EMG signals (sEMG) sampled at 800 Hz. A discrimination factor between physiological tremor (PH) and pathological tremor, namely, essential tremor (ET) and the tremor caused by Parkinson’s disease (PD), is obtained by summing the power entropy in band 6 (B6: 7.8125–9.375 Hz) and band 11 (B11: 15.625–17.1875 Hz). RESULTS: A discrimination accuracy of 93.87% is obtained between the PH group and the ET & PD group using a voting between three results obtained from the accelerometer signal and two sEMG signals. CONCLUSION: Biomedical signal processing techniques based on high resolution wavelet spectral analysis of accelerometer and sEMG signals are implemented to efficiently perform classification between physiological tremor and pathological tremor.

Pooling is a crucial aspect of Convolutional Neural Networks (CNNs), a prominent machine learning technique. It plays an essential role in the learning process by reducing the spatial dimensions of feature maps and minimizing computational costs. Common pooling methods, such as Max Pooling (MaxPool) and Average Pooling (AvgPool), have notable limitations. While MaxPool efficiently extracts important features, it often discards useful information, which can negatively impact CNN performance. In contrast, AvgPool preserves background information (maintaining global detail) but treats all inputs equally (assigning the same importance to every pixel in a region of an image), potentially leading to inefficiencies. We propose enhancing pooling adaptability by introducing learnable parameters to address these issues. These parameters (pooling kernels) are incorporated into the pooling methods to make them learnable. The study also explores how this modification affects the image classification performance of CNNs across various widely used datasets. The findings suggest that integrating parameters into the pooling process produces better results compared to the traditional pooling techniques previously discussed.

A soft decision algorithm for Obstructive Sleep Apnea (OSA) patient classification using R-R interval (RRI) data is investigated. This algorithm is based on fast and approximate estimation of the entropy of the wavelet-decomposed bands of the RRI dat

Atrial Fibrillation (AF) is a health-threatening condition, which is a violation of the heart rhythm that can lead to heart-related complications. Remarkable interest has been given to ECG signals analysis for AF detection in an early stage. In this context, we propose an artificial neural network ANN application to classify ECG signals into three classes, the first presents Normal Sinus Rhythm NSR, the second depicts abnormal signal with Atrial Fibrillation (AF) and the third shows noisy ECG signals. Accordingly, we achieve 93.1% accuracy classification results, 95.1% of sensitivity, 90.5% of specificity and 98%. Furthermore, we yield a value of zero error and a low value of cross entropy, which prove the robustness of the proposed ANN model architecture. Thus, we outperform the state of the art by achieving high accuracy classification without pre-processing step and without high level of feature extraction, and then we enable clinicians to determine automatically the class of each patient ECG signal.

Obstructive Sleep Apnea (OSA) is the cessation of breathing during sleep due to the collapse of upper airway. Polysomnographic recording is a conventional method for detection of OSA. Although it provides reliable results, it is expensive and cumbersome. Thus, an advanced non-invasive signal processing based technique is needed. The main purpose of this work is to predict the severity of sleep apnea using an efficient wavelet-based spectral analysis method of the heart rate variability (HRV) to classify sleep apnea into three different levels (mild, moderate, and severe) according to its severity and to distinguish them from normal subjects. The standard FFT spectrum analysis method and the soft-decision wavelet-based technique are to be used in this work in order to rank patients to full polysomnography. Data of 20 normal subjects and 20 patients with mild apnea and 20 patients with moderate apnea and 20 patients of severe apnea are used in this study. The data is obtained from the sleep laboratory of Sultan Qaboos University hospital in Oman. Four different classification versions have been used in this work. Accuracy result of 90% was obtained between severe and normal subjects and 85% between mild and normal and 75% between severe and moderate and 83.75% between normal and patients. The VLF/LF power spectral ratio of the wavelet-based soft-decision analysis of the RRI data after a high-pass filter resulted in the best accuracy of classification in all versions.