Computer Methods in Biomechanics and Biomedical Engineering最新文献

Complexities in processing human motion are possessed by lower limb exoskeletons. In this paper, a multi-task recognition model, IPTGNet, is proposed for the human locomotion modes. Temporal convolutional network and gated recurrent unit are parallelly fused through the dynamic tuning of hyperparameters using the improved particle swarm optimization algorithm. The experimental results demonstrate that faster and more stable convergence is achieved by IPTGNet with a recognition rate of 99.47% and a standard deviation of 0.42%. Furthermore, a finite state machine is utilized for incorrection of transition states. An innovative multi-task recognition of lower limb exoskeleton is provided by this paper.

PE is a serious form of pregnancy-related hypertension. Hypoxia can induce cellular dysfunction, adversely affecting both the infant and the mother. This study aims to investigate the relationship between HRGs and the diagnosis of PE, seeking to enhance our understanding of potential molecular mechanisms and offer new perspectives for the detection and treatment of the condition. A WGCNA network was established to identify key genes significantly associated with traits of PE. LASSO, SVM-RFE, and RF were utilized to identify feature genes. Calibration curves and DCA were employed to assess the diagnostic performance of the comprehensive nomogram. Consensus clustering was applied to identify subtypes of PE. GSEA and the construction of a ceRNA network were used to explore the potential biological functions and regulatory mechanisms of the identified feature genes. Furthermore, ssGSEA was conducted to investigate the immune landscape associated with PE. We successfully identified three potential diagnostic biomarkers for PE: P4HA1, NDRG1, and BHLHE40. Furthermore, the nomogram exhibited strong diagnostic performance. In patients with PE, the abundance of pro-inflammatory immune cells was significantly elevated, reflecting characteristics of high infiltration. The levels of immune cells infiltration were significantly correlated with the expression of the identified feature genes. Notably, these feature genes may be closely linked to mitochondrial-related biological functions. In conclusion, our findings enhance the understanding of the pathological mechanisms underlying PE and open innovative avenues for the diagnosis and treatment of PE.

This study enhances brain-controlled vehicle (BCV) lateral control using a steady-state visual evoked potential (SSVEP) interface and probabilistic support vector machine (SVM). A filter bank CSP (FBCSP) algorithm improves brain signal decoding, while a sigmoid-fitted SVM (SF-SVM) enables smoother control through probabilistic commands. Online tests achieved 84.03% classification accuracy. In lane-keeping tasks, SF-SVM improved completion rates by over 20% compared to standard SVM, reducing EEG non-stationarity effects. The probabilistic model optimized continuous control, significantly enhancing BCV performance.

Electroencephalogram (EEG)-based emotion recognition is a reliable and deployable method for identifying human emotional states. Currently, Graph convolution networks (GCN) have exhibited superior performance in extracting topological features of EEG. However, how to capture the dynamic topological relationship is still a challenge. In this paper, we propose an adaptive GCN with residual attention (AGC-RSTA) to extract the spatio-temporal discriminative features. Firstly, we construct an adaptive adjacency matrix in graph convolution, extracting the dynamic spatial topological features. We then utilize the residual spatio-temporal attention module to capture deep spatio-temporal features. Ablation studies and comparative experiments on the SEED and SEED-IV datasets demonstrate that our proposed model outperforms state-of-the-art methods, achieving recognition accuracies of 94.91% and 91.17%, respectively.

The present investigations are related to design a stochastic intelligent solver using the infrastructure of artificial neural networks (ANNs) and scaled conjugate gradient (SCG), i.e. ANNs-SCG for the numerical simulations of SIR dynamical prototype system based impacts of hospital bed. The SIR dynamical model is defined into three classes, susceptible patients in the hospital, infected population and recovered people. The proposed results are obtained through the sample statics of verification, testing and training of the dataset. The selection of the statics for training, testing and validation is chosen as 80%, 8% and 12%. A dataset is proposed based on the Adams scheme for the comparison of dynamical SIR prototype using the impacts of hospital bed. The numerical solutions are presented through the ANNs-SCG in order to reduce the values of the mean square error. To achieve the reliability, capability, accuracy, and competence of ANNs-SCG, the mathematical solutions are presented in the form of error histograms (EHs), regression, state transitions (STs) and correlation.

Currently, there is no established finite element (FE) method to apply physiologically realistic loads and constraints to the humerus. This FE study showed that 2 'simple' methods involving direct head loads, no head constraints, and rigid elbow or mid-length constraints created excessive stresses and bending. However, 2 'intermediate' methods involving direct head loads, but flexible head and elbow constraints, produced lower stresses and bending. Also, 2 'complex' methods involving muscles to generate head loads, plus flexible head and elbow constraints, generated the lowest stresses and moderate bending. This has implications for FE modeling research on intact and implanted humeri.

This article aims to improve and diversify signal processing techniques to execute a brain-computer interface (BCI) based on neurological phenomena observed when performing motor tasks using motor imagery (MI). The noise present in the original data, such as intermodulation noise, crosstalk, and other unwanted noise, is removed by Modify Least Mean Square (M-LMS) in the pre-processing stage. Traditional LMSs were unable to extract all the noise from the images. After pre-processing, the required features, such as statistical features, entropy features, etc., were extracted using Common Spatial Pattern (CSP) and Pearson's Correlation Coefficient (PCC) instead of the traditional single feature extraction model. The arithmetic optimization algorithm cannot select the features accurately and fails to reduce the feature dimensionality of the data. Thus, an Extended Arithmetic operation optimization (ExAo) algorithm is used to select the most significant attributes from the extracted features. The proposed model uses Double Attention Convolutional Neural Networks (DAttnConvNet) to classify the types of EEG signals based on optimal feature selection. Here, the attention mechanism is used to select and optimize the features to improve the classification accuracy and efficiency of the model. In EEG motor imagery datasets, the proposed model has been analyzed under class, which obtained an accuracy of 99.98% in class Baseline (B), 99.82% in class Imagined movement of a right fist (R) and 99.61% in class Imagined movement of both fists (RL). In the EEG dataset, the proposed model can obtain a high accuracy of 97.94% compared to EEG datasets of other models.

Motor imagery (MI) plays a crucial role in brain-computer interface (BCI), and the classification of MI tasks using electroencephalogram (EEG) is currently under extensive investigation. During MI classification, individual differences among subjects in terms of response and time latency need to be considered. Optimizing the time segment for different subjects can enhance subsequent classification performance. In view of the individual differences of subjects in motor imagery tasks, this article proposes a Time Segment Adaptive Optimization method based on Separability criterion and Correlation analysis (TSAOSC). The fundamental principle of this method involves applying the separability criterion to various sizes of time windows within the training data, identifying the optimal raw reference signal, and adaptively adjusting the time segment position for each trial's data by analyzing its relationship with the optimal reference signal. We evaluated our method on three BCI competition datasets, respectively. The utilization of the TSAOSC method in the experiments resulted in an enhancement of 4.90% in average classification accuracy compared to its absence. Additionally, building upon the TSAOSC approach, this study proposes a Nonlinear-TSAOSC method (N-TSAOSC) for analyzing EEG signals with nonlinearity, which shows improvements in the classification accuracy of certain subjects. The results of the experiments demonstrate that the proposed method is an effective time segment optimization method, and it can be integrated into other algorithms to further improve their accuracy.

This study analyzed the generation rule of streaming potential (SP) considering the piezoelectric effect of bone under static and dynamic compression. The piezoelectric equation was introduced into the equation of SP, and an osteon's 3D fluid-structure interaction finite element model with osteocytes was developed using COMSOL software. Seven working conditions helped study SP. The results showed positive and negative SP alternately acted on osteocyte under dynamic loads, and SP was about three orders of magnitude higher than that under static loads. Therefore, dynamic loads improved the osteocytes' force-electric microenvironment. The force-electric effect revealed the mechanism of treatment of osteoporosis.

This study presents the numerical solutions of the fractional schistosomiasis disease model (SDM) using the supervised neural networks (SNNs) and the computational scaled conjugate gradient (SCG), i.e. SNNs-SCG. The fractional derivatives are used for the precise outcomes of the fractional SDM. The preliminary fractional SDM is categorized as: uninfected, infected with schistosomiasis, recovered through infection, expose and susceptible to this virus. The accurateness of the SNNs-SCG is performed to solve three different scenarios based on the fractional SDM with synthetic data obtained with fractional Adams scheme (FAS). The generated data of FAS is used to execute SNNs-SCG scheme with 81% for training samples, 12% for testing and 7% for validation or authorization. The correctness of SNNs-SCG approach is perceived by the comparison with reference FAS results. The performances based on the error histograms (EHs), absolute error, MSE, regression, state transitions (STs) and correlation accomplish the accuracy, competence, and finesse of the SNNs-SCG scheme.