As stress related to many mental and physical health problems, monitoring stress and its management is getting increasingly important in modern societies. Because of the advantage of convolutional neural network (CNN) in automatic feature learning, this study proposed to use CNN to achieve accurate and fast detection of acute cognitive stress from heart rate variability (HRV). The traditional mental arithmetic calculation was adopted as the stressor for twenty participants, during which one-lead electrocardiogram (ECG) was acquired. Six conventional HRV methods for inferring cognitive stress were extracted from the ECG signals, and their performance in identifying acute cognitive stress was compared with the proposed CNN-based method. The experimental results showed that with a super-short (10 s) time window, the detection error rate of CNN was 17.3%, significantly better than the performance of all six conventional HRV methods (> 7.2%, p < 0.01). Further analysis showed that the improvement achieved by the proposed CNN methods mainly came from the decrease in false stress sample detection. This study demonstrated the possibility of super-short windows and the advantage of CNN on acute cognitive stress detection. Its outcome would benefit practical applications of real-time stress detection via HRV.

Muscle fatigue represents a complex physiological and psychological phenomenon that impairs physical performance and increases the risks of injury. It is important to continuously monitor fatigue levels for early detection and management of fatigue. The detection and classification of muscle fatigue also provide important information in human-computer interactions (HMI), sports injuries and performance, ergonomics, and prosthetic control. With this purpose in mind, this review first provides an overview of the mechanisms of muscle fatigue and its biomarkers and further enumerates various non-invasive techniques commonly used for muscle fatigue monitoring and detection in the literature, including electromyogram (EMG), which records the muscle electrical activity during muscle contractions, mechanomyogram (MMG), which records vibration signals of muscle fibers, near-infrared spectroscopy (NIRS), which measures the amount of oxygen in the muscle, ultrasound (US), which records signals of muscle deformation during muscle contractions. This review also introduces the principle and mechanism, parameters used for fatigue detection, application in fatigue detection, and advantages and disadvantages of each technology in detail. To conclude, the limitations/challenges that need to be addressed for future research in this area are presented.

Objective. This paper proposes machine learning models for mapping surface electromyography (sEMG) signals to regression of joint angle, joint velocity, joint acceleration, joint torque, and activation torque. Approach. The regression models, collectively known as MuscleNET, take one of four forms: ANN (forward artificial neural network), RNN (recurrent neural network), CNN (convolutional neural network), and RCNN (recurrent convolutional neural network). Inspired by conventional biomechanical muscle models, delayed kinematic signals were used along with sEMG signals as the machine learning model’s input; specifically, the CNN and RCNN were modeled with novel configurations for these input conditions. The models’ inputs contain either raw or filtered sEMG signals, which allowed evaluation of the filtering capabilities of the models. The models were trained using human experimental data and evaluated with different individual data. Main results. Results were compared in terms of regression error (using the root-mean-square) and model computation delay. The results indicate that the RNN (with filtered sEMG signals) and RCNN (with raw sEMG signals) models, both with delayed kinematic data, can extract underlying motor control information (such as joint activation torque or joint angle) from sEMG signals in pick-and-place tasks. The CNNs and RCNNs were able to filter raw sEMG signals. Significance. All forms of MuscleNET were found to map sEMG signals within 2 ms, fast enough for real-time applications such as the control of exoskeletons or active prostheses. The RNN model with filtered sEMG and delayed kinematic signals is particularly appropriate for applications in musculoskeletal simulation and biomechatronic device control.

Automatic gesture recognition (AGR) is investigated as an effortless human-machine interaction method, potentially applied in many industrial sectors. When using surface electromyogram (sEMG) for AGR, i.e. myoelectric control, a minimum of four EMG channels are required. However, in practical applications, fewer number of electrodes is always preferred, particularly for mobile and wearable applications. No published research focused on how to improve the performance of a myoelectric system with only two sEMG channels. In this study, we presented a systematic investigation to fill this gap. Specifically, we demonstrated that through spatial filtering and electrode position optimization, the myoelectric control performance was significantly improved (p < 0.05) and similar to that with four electrodes. Further, we found a significant correlation between offline and online performance metrics in the two-channel system, indicating that offline performance was transferable to online performance, highly relevant for algorithm development for sEMG-based AGR applications.

Recently, surface electromyography (sEMG) emerged as a novel biometric authentication method. Since EMG system parameters, such as the feature extraction methods and the number of channels, have been known to affect system performances, it is important to investigate these effects on the performance of the sEMG-based biometric system to determine optimal system parameters. In this study, three robust feature extraction methods, Time-domain (TD) feature, Frequency Division Technique (FDT), and Autoregressive (AR) feature, and their combinations were investigated while the number of channels varying from one to eight. For these system parameters, the performance of sixteen static wrist and hand gestures was systematically investigated in two authentication modes: verification and identification. The results from 24 participants showed that the TD features significantly ( ${p} < 0.05$ ) and consistently outperformed FDT and AR features for all channel numbers. The results also showed that the performance of a four-channel setup was not significantly different from those with higher number of channels. The average equal error rate (EER) for a four-channel sEMG verification system was 4% for TD features, 5.3% for FDT features, and 10% for AR features. For an identification system, the average Rank-1 error (R1E) for a four-channel configuration was 3% for TD features, 12.4% for FDT features, and 36.3% for AR features. The electrode position on the flexor carpi ulnaris (FCU) muscle had a critical contribution to the authentication performance. Thus, the combination of the TD feature set and a four-channel sEMG system with one of the electrodes positioned on the FCU are recommended for optimal authentication performance.

With the availability of high-density (HD) electrodes technology, the electrodes used in myoelectric control can have much higher density than the current practice. In this study, we investigated the effects of electrode density on pattern recognition (PR) based myoelectric control. Four density levels were analyzed in two directions: parallel and perpendicular to muscle fibers. Their influence on PR-based myoelectric control algorithms was investigated under three conditions between training and testing datasets: no electrode shift, 10-mm shift parallel to muscle fibers and 10-mm shift perpendicular to muscle fibers. The effect of electrode density varied among the different shift conditions: First, when there was no shift, increasing electrode density significantly improved the classification performance; second, when the shift was in the perpendicular direction, increasing electrode density resulted in deterioration in the classification performance; third, when the shift was in the parallel direction, the effect of the electrode density was more complicated-increasing the density in the parallel direction reduced the performance, while increasing density in the perpendicular direction would initially enhance the performance, but then reduce performance. To our best knowledge, this was the first study focusing on the role of electrode density in myoelectric control with the presence of electrode shift. Its outcome would benefit the design of electrode placement for future myoelectric prostheses with HD electrodes.