生物医学工程学杂志最新文献

Protein structure determines function, and structural information is critical for predicting protein thermostability. This study proposes a novel method for protein thermostability prediction by integrating graph embedding features and network topological features. By constructing residue interaction networks (RINs) to characterize protein structures, we calculated network topological features and utilize deep neural networks (DNN) to mine inherent characteristics. Using DeepWalk and Node2vec algorithms, we obtained node embeddings and extracted graph embedding features through a TopN strategy combined with bidirectional long short-term memory (BiLSTM) networks. Additionally, we introduced the Doc2vec algorithm to replace the Word2vec module in graph embedding algorithms, generating graph embedding feature vector encodings. By employing an attention mechanism to fuse graph embedding features with network topological features, we constructed a high-precision prediction model, achieving 87.85% prediction accuracy on a bacterial protein dataset. Furthermore, we analyzed the differences in the contributions of network topological features in the model and the differences among various graph embedding methods, and found that the combination of DeepWalk features with Doc2vec and all topological features was crucial for the identification of thermostable proteins. This study provides a practical and effective new method for protein thermostability prediction, and at the same time offers theoretical guidance for exploring protein diversity, discovering new thermostable proteins, and the intelligent modification of mesophilic proteins.

Although metal blocks have been widely used for reconstructing uncontained tibial bone defects, the influence of their elastic modulus on the stability of tibial prosthesis fixation remains unclear. Based on this, a finite element model incorporating constrained condylar knee (CCK) prosthesis, tibia, and metal block was established. Considering the influence of the post-restraint structure of the prosthesis, the effects of variations in the elastic modulus of the block on the von Mises stress distribution in the tibia and the block, as well as on the micromotion at the bone-prosthesis fixation interface, were investigated. Results demonstrated that collision between the insert post and femoral prosthesis during tibial internal rotation increased tibial von Mises stress, significantly influencing the prediction of block elastic modulus variation. A decrease in the elastic modulus of the metal block resulted in increased von Mises stress in the proximal tibia, significantly reduced von Mises stress in the distal tibia, decreased von Mises stress of the block, and increased micromotion at the bone-prosthesis fixation interface. When the elastic modulus of the metal block fell below that of bone cement, inadequate block support substantially increased the risk of stress shielding in the distal tibia and fixation interface loosening. Therefore, this study recommends that biomechanical investigations of CCK prostheses must consider the post-constraint effect, and the elastic modulus of metal blocks for bone reconstruction should not be lower than 3 600 MPa.

The existing epilepsy seizure detection algorithms have problems such as overfitting and poor generalization ability due to high reliance on manual labeling of electroencephalogram's data and data imbalance between seizure and interictal periods. An unsupervised learning detection method for epileptic seizure that jointed graph attention network (GAT) and Transformer framework (GAT-T) was proposed. In this method, channel correlations were adaptively learned by GAT encoder. Temporal information was captured by one-dimensional convolution decoder. Combining outputs of the two mentioned above, predicted values for electroencephalogram were generated. The collective anomaly score was calculated and the detection threshold was determined. The results demonstrated that GAT-T achieved the average performance exceeding 90% (or 99%) with a 0.25 s (or 2 s) time segment length, which could effectively detect epileptic seizures. Moreover, the channel association probability matrix was expected to assist clinicians in the initial screening of the epileptogenic zone, and ablation experiments also reflected the significance of each module in GAT-T. This study may assist clinicians in making more accurate diagnostic and therapeutic decisions for epilepsy patients.

Exercise intervention is an important non-pharmacological intervention for various diseases, and establishing precise exercise load assessment techniques can improve the quality of exercise intervention and the efficiency of disease prevention and control. Based on data collection from wearable devices, this study conducts nonlinear optimization and empirical verification of the original "Fitness-Fatigue Model". By constructing a time-varying attenuation function and specific coefficients, this study develops an optimized mathematical model that reflects the nonlinear characteristics of training responses. Thirteen participants underwent 12 weeks of moderate-intensity continuous cycling, three times per week. For each training session, external load (actual work done) and internal load (heart rate variability index) data were collected for each individual to conduct a performance comparison between the optimized model and the original model. The results show that the optimized model demonstrates a significantly improved overall goodness of fit and superior predictive ability. In summary, the findings of this study can support dynamic adjustments to participants' training programs and aid in the prevention and control of chronic diseases.

In recent years, the ongoing development of transcranial electrical stimulation (TES) and transcranial magnetic stimulation (TMS) has demonstrated significant potential in the treatment and rehabilitation of various brain diseases. In particular, the combined application of TES and TMS has shown considerable clinical value due to their potential synergistic effects. This paper first systematically reviews the mechanisms underlying TES and TMS, highlighting their respective advantages and limitations. Subsequently, the potential mechanisms of transcranial electromagnetic combined stimulation are explored, with a particular focus on three combined stimulation protocols: Repetitive TMS (rTMS) with transcranial direct current stimulation (tDCS), rTMS with transcranial alternating current stimulation (tACS), and theta burst TMS (TBS) with tACS, as well as their clinical applications in brain diseases. Finally, the paper analyzes the key challenges in transcranial electromagnetic combined stimulation research and outlines its future development directions. The aim of this paper is to provide a reference for the optimization and application of transcranial electromagnetic combined stimulation schemes in the treatment and rehabilitation of brain diseases.

In order to accurately capture the respiratory muscle movement and extract the synchronization signals corresponding to the breathing phases, a comprehensive signal sensing system for sensing the movement of the respiratory muscle was developed with applying the thin-film varistor FSR402 IMS-C07A in this paper. The system integrated a sensor, a signal processing circuit, and an application program to collect, amplify and denoise electronic signals. Based on the respiratory muscle movement sensor and a STM32F107 development board, an experimental platform was designed to conduct experiments. The respiratory muscle movement data and respiratory airflow data were collected from 3 healthy adults for comparative analysis. In this paper, the results demonstrated that the method for determining respiratory phase based on the sensing the respiratory muscle movement exhibited strong real-time performance. Compared to traditional airflow-based respiratory phase detection, the proposed method showed a lead times ranging from 33 to 210 ms [(88.3 ± 47.9) ms] for expiration switched into inspiration and 17 to 222 ms [(92.9 ± 63.8) ms] for inspiration switched into expiration, respectively. When this system is applied to trigger the output of the ventilator, it will effectively improve the patient-ventilator synchrony and facilitate the ventilation treatment for patients with respiratory diseases.

Electroencephalogram (EEG) serves as an effective indicator of detecting fatigue driving. Utilizing the open accessible Shanghai Jiao Tong University Emotion Electroencephalography Dataset (SEED-VIG), driving states are divided into three categories including awake, tired and drowsy for investigation. Given the characteristics of mutual influence and interdependence among EEG channels, as well as the consistency of the graph convolutional neural network (GCNN) structure, we designed an adjacency matrix based on the Pearson correlation coefficients of EEG signals among channels and their positional relationships. Subsequently, we developed a GCNN for recognition. The experimental results show that the average classification accuracy of driving state categories for 20 subjects, from the SEED-VIG dataset under the smooth feature of differential entropy (DE) linear dynamic system is 91.66%. Moreover, the highest classification accuracy can reach 98.87%, and the average Kappa coefficient is 0.83. This work demonstrates the reliability of this method and provides a guideline for the research field of safe driving brain computer interface.

Brain-computer interface (BCI) technology faces structural risks due to a misalignment between its technological maturity and industrialization expectations. This study used the Technology Readiness Level (TRL) framework to assess the status of major BCI paradigms-such as steady-state visual evoked potential (SSVEP), motor imagery, and P300-and found that they predominantly remained at TRL4 to TRL6, with few stable applications reaching TRL9. The analysis identified four interrelated sources of bubble risk: overly broad definitions of BCI, excessive focus on decoding performance, asynchronous translational progress, and imprecise terminology usage. These distortions have contributed to the misallocation of research resources and public misunderstanding. To foster the sustainable development of BCI, this paper advocated the establishment of a standardized TRL evaluation system, clearer terminological boundaries, stronger support for fundamental research, enhanced ethical oversight, and the implementation of inclusive and diversified governance mechanisms.

Small animal multimodal biomedical imaging refers to the integration of multiple imaging techniques within the same system or device to acquire comprehensive physiological and pathological information of small animals, such as mice and rats. With the continuous advancement of biomedical research, this cutting-edge technology has attracted extensive attention. Multimodal imaging techniques, based on diverse imaging principles, overcome the limitations of single-modal imaging through information fusion, significantly enhancing the overall system's sensitivity, temporal/spatial resolution, and quantitative accuracy. In the future, the integration of new materials and artificial intelligence will further boost its sensitivity and resolution. Through interdisciplinary innovation, this technology is expected to become the core technology of personalized medicine and expand its applications to drug development, environmental monitoring, and other fields, thus reshaping the landscape of biomedical research and clinical practice. This review summarized the progress on the application and investigation of multimodal biomedical imaging techniques, and discussed its development in the future.

Brain age prediction, as a significant approach for assessing brain health and early diagnosing neurodegenerative diseases, has garnered widespread attention in recent years. Electroencephalogram (EEG), an non-invasive, convenient, and cost-effective neurophysiological signal, offers unique advantages for brain age prediction due to its high temporal resolution and strong correlation with brain functional states. Despite substantial progress in enhancing prediction accuracy and generalizability, challenges remain in data quality and model interpretability. This review comprehensively examined the advancements in EEG-based brain age prediction, detailing key aspects of data preprocessing, feature extraction, model construction, and result evaluation. It also summarized the current applications of machine learning and deep learning methods in this field, analyzed existing issues, and explored future directions to promote the widespread application of EEG-based brain age prediction in both clinical and research settings.