Ankle rehabilitation robots represent an important branch of rehabilitation robotics, offering significant potential to improve the quality of life for patients with ankle dysfunction caused by stroke, sports injuries, and other conditions. This review first outlines the anatomy and range of motion of the ankle joint, compares conventional rehabilitation approaches with robot-assisted therapy, and highlights the clinical significance of ankle rehabilitation robots. It then systematically examines current research progress from two core perspectives: mechanical structure design and control strategies. In mechanical design, the performance characteristics of series versus parallel mechanisms are compared, the advantages and limitations of actuation methods such as electric motors and pneumatic artificial muscles are analyzed, and the application contexts of platform-based and wearable robots are discussed. In control strategies, the discussion covers motion control and human-robot interaction, beginning with fundamental position, velocity, and trajectory tracking control, and extending to intention-level and cognitive interaction. Finally, based on current research and clinical needs, future ankle rehabilitation robots are expected to evolve toward greater flexibility, intelligence, and universality, providing a theoretical foundation for future studies.

Objective: This study aimed to apply machine learning approaches to the Kellgren-Lawrence (KL) grading of knee osteoarthritis, develop an effective automatic KL grading technique, and provide a methodological reference for clinical diagnosis and research. Methods: Data were obtained from the Osteoarthritis Initiative (OAI) knee X-ray image dataset, comprising 8,110 images from the folders of auto_test, train, and val. All images were first subjected to inversion processing, followed by extraction of two-dimensional radiomic features. Feature selection was then conducted using a combination of variance thresholding and analysis of variance (ANOVA), yielding 18 key features. To address class imbalance in the original dataset, this synthetic minority over-sampling technique (SMOTE)  and class weight balancing were jointly applied. Eight machine learning models-Decision Trees (DT), Logistic Regression (LR), Random Forests (RF), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and Light Gradient Boosting Machine (LightGBM)-were trained for KL grading of knee osteoarthritis. Model performance was evaluated using accuracy, precision, recall, F1-score, and the area under the curve (AUC). For the optimal SVM model, global and local interpretability analy-ses were further conducted using SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) to identify the key factors influencing model decisions. Results: The support vector machine  (SVM) model achieves the best performance. Conclusion: This study establishes an effective machine learningbased method for automatic KL grading of knee osteoarthritis, providing valuable support for clinical diagnosis and research applications.

Objective: To design a testing device for measuring the radial support force and bending spring back force of stent grafts and evaluate its effectiveness. Methods: A radial force and spring-back force testing device was designed to integrate with a tensile testing machine. The radial compression and bending characteristics of stent grafts for thoracic aorta applications were analyzed, and the corresponding conversion formulas were derived. A custom stent ring fixture was fabricated, and a five-wave gradient stent was sewn. Both physical experiments and finite element simulations were conducted. Radial support forces were measured by gripping 20%, 40%, and 60% of the stent’s diameter, and bending tests were performed at angles of 60°, 90°, and 180°. The stability of the testing device was analyzed through comparative tests across different compression diameters and bending angles. Results: The device demonstrated high detection precision, stability, and accuracy, with minimal deviation across multiple measurements. The mechanical behavior of the stent observed in both finite element simulations and physical experiments showed consistent results. Conclusions: The testing device developed in this study effectively measures the mechanical changes in large-diameter stent grafts, providing a new reference for testing large-diameter stents.

To enhance the adaptive learning, self-organization, and fault tolerance capabilities of gait pattern recognition in intelligent knee prostheses, an online walking pattern recognition method based on the convolutional neural networks (CNN)-long short term memory (LSTM) model is proposed. Five test subjects wore the intelligent knee prostheses and performed four walking modes: level walking, uphill walking, downhill walking, and stair descent. The preprocessed gait data were fed into four neural network models: CNN, LSTM, CNN-LSTM, and CNN-bidirectional LSTM. Through hyperparameter tuning, the recognition accuracy of these models was compared. Real-time indicator, gait recognition delay, was also measured. Experimental results showed each model had its strengths and weaknesses. Overall, the CNN-LSTM model achieved the best recognition performance with accuracy rates of: level walking 89%±2.5%, uphill 72.8%±3.2%, downhill 71%±3.2%, and stair descent 96%±2.5%. When switching from level walking to downhill, gait recognition delay was 51.7%±15.6%, and vice versa it was 75.8%±11.5%; when switching from level walking to stair descent, gait recognition delay was 47.1%±17.1%, and vice versa it was 38.6%±10.5%. In summary, the application of the CNN-LSTM model for walking pattern recognition in unilateral intelligent knee prostheses is feasible, with accuracy and real-time performance meeting the control requirements of the prostheses.

The growing demand for healthcare has brought Traditional Chinese Medicine (TCM) massage therapy into the spotlight in academic circles. Numerous studies have underscored the effectiveness of TCM massage in health promotion, disease amelioration, and quality of life enhancement. However, the field faces challenges such as inconsistent training and inadequate transfer of experiential knowledge. Recently, machine learning has shown potential in the medical field and its application in TCM massage therapy offers new developmental opportunities. This paper reviews key research areas exploring the synergy between machine learning and Chinese massage therapy, including acupoint localization and identification, massage practice, and personalized treatment plans. It summarizes progress and identifies the challenges in integrating these technologies. Despite potential risks, merging these technologies is poised to be a trend in future healthcare, driven by advances in computer technology and the needs of TCM practitioners.

Spike sorting plays a pivotal role in neuroscience, serving as a crucial step of separating electrical signals recorded from multiple neurons to further analyze neuronal interactions. This process involves separating electrical signals that originate from multiple neurons, recorded through devices like electrode arrays. This is a very important  link in the field of brain-computer interfaces. The objective of spike sorting algorithm (SSA) is to distinguish the  behavior of one or more neurons from background noise using the waveforms captured by brain-embedded electrodes. This article starts from the steps of the conventional SSA and divides the SSA into three steps: spike detection, spike feature extraction, and spike clustering. It outlines prevalent algorithms for each phase before delving  into two emerging technologies: template matching and deep learning-based methods. The discussion on deep  learning is further subdivided into three approaches: end-to-end solution, deep learning for spike sorting steps,  and spiking neural networks-based solutions. Finally, it elaborates future challenges and development trends of SSAs.

The ability to interpret vibration signals in the biomedical field offers a promising path toward continuous improvement of medical devices. By examining the revolutions per minute profile, analysts can identify any deviations or anomalies in the vibration patterns at different speeds. This information can help identify potential faults or imbalances within the rotating machinery. With a comprehensive understanding of the revolutions per minute profile, analysts can make informed decisions regarding maintenance and repairs. Besides, the analysis of the order of vibration signals represents an essential pillar of biomedical engineering, bringing an innovative and in-depth perspective to the development of medical devices, and contributing to the continued advancement of medical technology and healthcare. Integrating vibration analysis into preventive maintenance practices can help ensure the reliability of medical equipment, reduce potential risks to patients, and contribute to the advancement of healthcare quality. 

Accurately and efficiently detecting the quantity of blood cells is crucial in routine blood examination, as abnormal high or low numbers of blood cells are associated with the occurrence of various disorders. Due to inherent drawbacks, traditional blood cell analysis equipment cannot meet the demands of modern primary healthcare, particularly in terms of point-of-care testing. In recent years, the development of point-of-care testing blood cell counting equipment has been accelerated, thanks to the rapid advancement of microfluidic technology and the expanding research on blood cell counting using microfluidic chips. In this paper, we reviewed three blood cell counting methods based on microfluidic chips, electrical impedance, light scattering, and microscopic imaging, as well as the recent development and achievements in blood cell counting using microfluidic chips.

Objective: To explore the correlation among paraspinal muscle functions electromyography (EMG), muscle stiffness, and pain threshold in patients with adolescent idiopathic scoliosis (AIS). Methods: Eighteen patients with AIS were recruited. A Noraxon system equipped with four wireless EMG sensors was used to collect EMG data on the paraspinal muscles in relaxed standing and weight-bearing standing states. Muscle stiffness and pain threshold were measured using a muscle tonometer. The differences in mean EMG amplitude, muscle stiffness, and pain threshold between the concave and convex sides of the scoliosis were analyzed. Results: Among patients with different scoliosis locations, Cobb angles, ages, and brace treatment durations, the mean EMG amplitude of the paraspinal muscles on the convex side of scoliosis was significantly higher than that on the concave side (P<0.05). The muscle stiffness and pain threshold of the paraspinal muscles on the convex side were also significantly higher than those on the concave side (both P<0.05). There was a low correlation between the mean EMG amplitude of the paraspinal muscles, muscle stiffness, and pain threshold (R<0.5, P>0.05). Conclusion: In AIS patients, the electromyographic activity, muscle stiffness, and pain threshold of the paraspinal muscles on the convex side of scoliosis were all higher than those on the concave side, and the correlation among the three indicators was low.