Background: Mental fatigue is usually caused by long-term cognitive activities, mainly manifested as drowsiness, difficulty in concentrating, decreased alertness, disordered thinking, slow reaction, lethargy, reduced work efficiency, error-prone and so on. Mental fatigue has become a widespread sub-health condition, and has a serious impact on the cognitive function of the brain. However, seldom studies investigate the differences of mental fatigue on electrophysiological activity both in resting state and task state at the same time. Here, twenty healthy male participants were recruited to do a consecutive mental arithmetic tasks for mental fatigue induction, and electroencephalogram (EEG) data were collected before and after each tasks. The power and relative power of five EEG rhythms both in resting state and task state were analyzed statistically. Results: The results of brain topographies and statistical analysis indicated that mental arithmetic task can successfully induce mental fatigue in the enrolled subjects. The relative power index was more sensitive than the power index in response to mental fatigue, and the relative power for assessing mental fatigue was better in resting state than in task state. Furthermore, we found that it is of great physiological significance to divide alpha frequency band into alpha1 band and alpha2 band in fatigue related studies, and at the same time improve the statistical differences of sub-bands. Conclusions: Our current results suggested that the brain activity in mental fatigue state has great differences in resting state and task state, and it is imperative to select the appropriate state in EEG data acquisition and divide alpha band into alpha1 and alpha2 bands in mental fatigue related researches.

Brain functional network has been widely applied to investigate brain function changes among different conditions and proved to be a small-world-like network. But seldom researches explore the effects of mental fatigue on the small-world brain functional network organization. In the present study, 20 healthy individuals were included to do a consecutive mental arithmetic task to induce mental fatigue, and scalp electroencephalogram (EEG) signals were recorded before and after the task. Correlations between all pairs of EEG channels were determined by mutual information (MI). The resulting adjacency matrices were converted into brain functional networks by applying a threshold, and then, the clustering coefficient (C), characteristic path length (L), and corresponding small-world feature were calculated. Through performing analysis of variance (ANOVA) on the mean MI for every EEG rhythm, only the data of α1 rhythm during the task state were emerged for the further explorations of mental fatigue. For a wide range of thresholds, C increased and L and small-world feature decreased with the deepening mental fatigue. The pattern of the small-world characteristic still existed when computed with a constant degree. Our present findings indicated that more functional connectivities were activated at the mental fatigue stage for efficient information transmission and processing, and mental fatigue can be characterized by a reduced small-world network characteristic. Our results provide a new perspective to understand the neural mechanisms of mental fatigue based on complex network theories.

Mechanical signals are not only disturbed by Gaussian noise, but also by non-Gaussian noise. These Gaussian noise and non-Gaussian noise have gravely impeded detecting of rolling bearing defects using traditional methods. In this context, the paper develops a novel detection method for rolling bearing, which combines bispectrum analysis with an improved ensemble empirical mode decomposition (EEMD). To effectively eliminate Gaussian noise in the signal, bispectrum analysis is adopted. In order to effectively reduce non-Gaussian noise, a cloud model-improved EEMD is proposed, where the cloud model is introduced to restrain the mode mixing phenomenon. Then a rolling bearing defect detection plan based on the proposed method is put forward. From theoretical analysis and experimental verification, it is demonstrated that the proposed method has superior performance in reducing multiple background noise. Furthermore, compared with other three methods, the results show that the proposed method can detect the defect of rolling bearings more effectively.

Rolling element bearings are crucial, of high failure rate and easy damage parts of rotating machinery, and significantly affect safe and reliable production processes. Much more attention has been focused on fault diagnosis of rolling element bearings in recent years. This article presents a novel feature extraction scheme for the classification of multiple bearing faults. Multi-Scale Sample Entropy (M-SSampEn) is combined with Energy Moment (EM) to construct a time-domain Multi-Scale Sample Entropy-based Energy Moment (M-SSampEn-EM) feature extractor. The classifier model for the proposed fault classification system has been built using the Least Square Support Vector Machine (LS-SVM). The M-SSampEn-EM feature extractor is used to capture two-dimensional representative eigenvectors from multiple fault classes' bearing vibration data. Its separability performance is ensured by optimizing the feature-extraction parameters, including the Embedded Dimension and the Tolerance, etc. The LS-SVM classifier is also compared with two Neural Network (NN)-based classifiers, i.e., Radial Basis Function NN (RBFNN) and Probabilistic NN (PNN), to show it better generalization performance on bearing fault classification. The experimental study verifies the excellent capacity of the proposed approach in bearing fault classification.

Driving fatigue is the main cause of traffic accidents, which seriously affects people’s life and property safety. Many researchers have applied electroencephalogram (EEG) signals for driving fatigue detection to reduce negative effects. The main challenges are the practicality and accuracy of the EEG-based driving fatigue detection method when it is applied on the real road. In our previous study, we attempted to improve the practicality of fatigue detection based on the proposed non-hair-bearing (NHB) montage with fewer EEG channels, but the recognition accuracy was only 76.47% with the random forest (RF) model. In order to improve the accuracy with NHB montage, this study proposed an improved transformer architecture for one-dimensional feature vector classification based on introducing the Gated Linear Unit (GLU) in the Attention sub-block and Feed-Forward Networks (FFN) sub-block of a transformer, called GLU-Oneformer. Moreover, we constructed an NHB-EEG-based feature set, including the same EEG features (power ratio, approximate entropy, and mutual information (MI)) in our previous study, and the lateralization features of the power ratio and approximate entropy based on the strategy of brain lateralization. The results indicated that our GLU-Oneformer method significantly improved the recognition performance and achieved an accuracy of 86.97%. Our framework demonstrated that the combination of the NHB montage and the proposed GLU-Oneformer model could well support driving fatigue detection.

To address the poor performance of traditional rolling bearing fault diagnosis models in cross-condition tasks due to significant feature differences, a Transitional Domain Adversarial Network (TDAN) proposed in this paper. This model initially builds a multi-channel, multi-feature extractor to obtain the frequency domain phase spectrum of vibration signals. It then integrates this data with spectral and time-domain features to extract deep, domain-invariant characteristics from various perspectives. Transition units are also designed to derive both domain and class transitional zones. The domain transitional zone aims to mitigate the loss of certain features caused by forced alignment between source and target domains. Meanwhile, the class transitional zone enhances feature granularity from the perspective of inter-class variation, thereby improving class-specific representation, smoothing the adversarial process, and boosting model generalization. Additionally, to address the target-oriented adversarial loss function, a re-adversarial module is introduced. This process equips the model with the capability to escape local optima and optimize parameters adaptively during training, resulting in stronger robustness and adaptability. Comparative experiments with other unsupervised domain adaptation (UDA) methods on two bearing datasets demonstrate TDAN’s effectiveness and superiority in rolling bearing cross-condition fault diagnosis. It also demonstrates the model’s potential for application in real industrial scenarios where varying operating conditions lead to differences in vibration signals.