Robotic lower-limb rehabilitation training is a better alternative for the physical training efforts of a therapist due to advantages, such as intensive repetitive motions, economical therapy, and quantitative assessment of the level of motor recovery through the measurement of force and movement patterns. However, in actual robotic rehabilitation training, emergency stops occur frequently to prevent injury to patients. However, frequent stopping is a waste of time and resources of both therapists and patients. Therefore, early detection of emergency stops in real-time is essential to take appropriate actions. In this paper, we propose a novel deep-learning-based technique for detecting emergency stops as early as possible. First, a bidirectional long short-term memory prediction model was trained using only the normal joint data collected from a real robotic training system. Next, a real-time threshold-based algorithm was developed with cumulative error. The experimental results revealed a precision of 0.94, recall of 0.93, and F1 score of 0.93. Additionally, it was observed that the prediction model was robust for variations in measurement noise.

The timed up-and-go (TUG) test is an efficient way to evaluate an individual’s basic functional mobility, such as standing up, walking, turning around, and sitting back. The total completion time of the TUG test is a metric indicating an individual’s overall mobility. Moreover, the fine-grained consumption time of the individual subtasks in the TUG test may provide important clinical information, such as elapsed time and speed of each TUG subtask, which may not only assist professionals in clinical interventions but also distinguish the functional recovery of patients. To perform more accurate, efficient, robust, and objective tests, this paper proposes a novel deep learning-based subtask segmentation of the TUG test using a dilated temporal convolutional network with a single RGB-D camera. Evaluation with three different subject groups (healthy young, healthy adult, stroke patients) showed that the proposed method demonstrated better generality and achieved a significantly higher and more robust performance (healthy young = 95.458%, healthy adult = 94.525%, stroke = 93.578%) than the existing rule-based and artificial neural network-based subtask segmentation methods. Additionally, the results indicated that the input from the pelvis alone achieved the best accuracy among many other single inputs or combinations of inputs, which allows a real-time inference (approximately 15 Hz) in edge devices, such as smartphones.

Robot-assisted gait training (RAGT) is a promising therapeutic vehicle to maximize active participation and enhance functional neuroplasticity in patients with central nervous system pathology by adequately adjusting gait speed, body weight support (BWS) level, and impedance provided by the exoskeleton. The aim of the present study was to determine the relationship between RAGT training parameters (BWS and speed) and electromyography (EMG) muscle activity torques in the knee and hip joint during RAGT. To analyze the correlation between the joint torques measured in the Walkbot gait rehabilitation system and the EMG signal of the lower limbs (vastus lateralis oblique, biceps femoris, tibialis anterior, and gastrocnemius) and understand the real-time state of the lower limb an experiment involving 20 subjects was conducted. The EMG–torque relationship was evaluated in a general rehabilitation training setting to overcome the limitations of in vivo settings. Pearson correlation coefficient analysis was performed at p < 0.05. Moderate relationships between biceps femoris activation data and hip and knee torques were statistically significant, ranging from r = 0.412 to −0.590, p < 0.05). Importantly, inverse relationships existed between hip torques and vastus lateralis oblique, biceps femoris, and tibialis anterior activation, respectively. The present results demonstrated the association between EMG locomotor control patterns and torque generation in the hip and knee joints during RAGT-treadmill under the different BWS and walking speed settings while adjusting the impedance mode parameters in non-neurological adults. Additionally, the EMG locomotor control patterns, concurrent torque generation in the hip and knee joints, and application of different BWS and walking speed parameters in the RAGT were linked to the gait speed and BWS. The outcomes also showed that the amount of BWS supplied had an impact on the effects of treadmill speed on muscle activity and temporal step control. It is essential to adjust RAGT parameters precisely in order to maximize training session efficiency and quality. The results of this study nevertheless call for more investigation into the relationship between muscle activity and torque outcomes in diseased populations with gait impairment.

For people with knee osteoarthritis (OA), step-up is one of the most physically demanding and pain-evoking daily movements. Given the significant role of the prefrontal cortex (PFC) in executive function and pain modulation, our objectives were (1) to assess PFC activation during walking and step-up in people with knee OA and (2) to examine the association of pain and psychological outcomes with changes in PFC activation from walking to step-up. This cross-sectional study included 44 individuals with symptomatic knee OA. Participants completed Knee Injury and Osteoarthritis Outcome Score (KOOS), Pain, Enjoyment of Life, and General Activity (PEG) scale, Patient Health Questionnaire (PHQ-8) for mood, 3-item Pain Catastrophizing Scale (PCS), and Fear-Avoidance Beliefs Questionnaire (FABQ). Oxygenated hemoglobin concentration changes (HbO₂) in the bilateral PFC were assessed using functional Near-Infrared Spectroscopy. HbO₂ were compared between tasks using repeated measures ANCOVA. Multiple linear regression models tested the association between clinical outcomes and changes in HbO₂ from walking to step-up while adjusting for confounders. HbO₂ during step-up was higher compared to walking, with moderate effect size (d ​= ​0.3-0.5). Greater pain catastrophizing (R² ​= ​0.10) was correlated with smaller changes in HbO₂ from walking to step-up. People with knee OA utilized greater executive control during step-up compared to walking. Similar to stair climbing, step-up is more physically demanding and more frequently impaired than walking. Greater pain catastrophizing was related to smaller increases in PFC activation from walking to step-up, reflecting the availability of fewer executive resources to maintain task performance.