Tool-sensed object information effectively supports vision for multisensory grasping.

Tools enable humans to extend their sensing abilities beyond the natural limits of their hands, allowing them to sense objects as if they were using their hands directly. The similarities between direct hand interactions with objects (hand-based sensing) and the ability to extend sensory information processing beyond the hand (tool-mediated sensing) entail the existence of comparable processes for integrating tool- and hand-sensed information with vision, raising the question of whether tools support vision in bimanual object manipulations. Here, we investigated participants' performance while grasping objects either held with a tool or with their hand and compared these conditions with visually guided grasping (Experiment 1). By measuring reaction time, peak velocity, and peak of grip aperture, we found that actions were initiated earlier and performed with a smaller peak grip aperture when the object was seen and held with the tool or the contralateral hand compared to when it was only seen. Thus, tool-mediated sensing effectively supports vision in multisensory grasping and, even more intriguingly, resembles hand-based sensing. We excluded that results were due to the force exerted on the tool's handle (Experiment 2). Additionally, as for hand-based sensing, we found evidence that the tool supports vision by mainly providing object positional information (Experiment 3). Thus, integrating the tool-sensed position of the object with vision is sufficient to promote a multisensory advantage in grasping. Our findings indicate that multisensory integration mechanisms significantly improve grasping actions and fine-tune contralateral hand movements even when object information is only indirectly sensed through a tool. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Student Motivation Analysis Based on Raising-Hand Videos.

In current smart classroom research, numerous studies focus on recognizing hand-raising, but few analyze the movements to interpret students' intentions. This limitation hinders teachers from utilizing this information to enhance the effectiveness of smart classroom teaching. Assistive teaching methods, including robotic and artificial intelligence teaching, require smart classroom systems to both recognize and thoroughly analyze hand-raising movements. This detailed analysis enables systems to provide targeted guidance based on students' hand-raising behavior. This study proposes a morphology-based analysis method to innovatively convert students' skeleton key point data into several one-dimensional time series. By analyzing these time series, this method offers a more detailed analysis of student hand-raising behavior, addressing the limitations of deep learning methods that cannot compare classroom hand-raising enthusiasm or establish a detailed database of such behavior. This method primarily utilizes a neural network to obtain students' skeleton estimation results, which are then converted into time series of several variables using the morphology-based analysis method. The YOLOX and HrNet models were employed to obtain the skeleton estimation results; YOLOX is an object detection model, while HrNet is a skeleton estimation model. This method successfully recognizes hand-raising actions and provides a detailed analysis of their speed and amplitude, effectively supplementing the coarse recognition capabilities of neural networks. The effectiveness of this method has been validated through experiments.

Positional Analysis of Assisting Muscles for Handling-Assisted Exoskeletons.

In order to better design handling-assisted exoskeletons, it is necessary to analyze the biomechanics of human hand movements. In this study, Anybody Modeling System (AMS) simulation was used to analyze the movement state of muscles during human handling. Combined with surface electromyography (sEMG) experiments, specific analysis and verification were carried out to obtain the position of muscles that the human body needs to assist during handling. In this study, the simulation and experiment were carried out for the manual handling process. A treatment group and an experimental group were set up. This study found that the vastus medialis muscle, vastus lateralis muscle, latissimus dorsi muscle, trapezius muscle, deltoid muscle and triceps brachii muscle require more energy in the process of handling, and it is reasonable and effective to combine sEMG signals with the simulation of the musculoskeletal model to analyze the muscle condition of human movement.

An Intelligent Hand-Assisted Diagnosis System Based on Information Fusion.

This research proposes an innovative, intelligent hand-assisted diagnostic system aiming to achieve a comprehensive assessment of hand function through information fusion technology. Based on the single-vision algorithm we designed, the system can perceive and analyze the morphology and motion posture of the patient's hands in real time. This visual perception can provide an objective data foundation and capture the continuous changes in the patient's hand movement, thereby providing more detailed information for the assessment and providing a scientific basis for subsequent treatment plans. By introducing medical knowledge graph technology, the system integrates and analyzes medical knowledge information and combines it with a voice question-answering system, allowing patients to communicate and obtain information effectively even with limited hand function. Voice question-answering, as a subjective and convenient interaction method, greatly improves the interactivity and communication efficiency between patients and the system. In conclusion, this system holds immense potential as a highly efficient and accurate hand-assisted assessment tool, delivering enhanced diagnostic services and rehabilitation support for patients.

Stretchable glove for accurate and robust hand pose reconstruction based on comprehensive motion data.

We propose a compact wearable glove capable of estimating both the finger bone lengths and the joint angles of the wearer with a simple stretch-based sensing mechanism. The soft sensing glove is designed to easily stretch and to be one-size-fits-all, both measuring the size of the hand and estimating the finger joint motions of the thumb, index, and middle fingers. The system was calibrated and evaluated using comprehensive hand motion data that reflect the extensive range of natural human hand motions and various anatomical structures. The data were collected with a custom motion-capture setup and transformed into the joint angles through our post-processing method. The glove system is capable of reconstructing arbitrary and even unconventional hand poses with accuracy and robustness, confirmed by evaluations on the estimation of bone lengths (mean error: 2.1 mm), joint angles (mean error: 4.16°), and fingertip positions (mean 3D error: 4.02 mm), and on overall hand pose reconstructions in various applications. The proposed glove allows us to take advantage of the dexterity of the human hand with potential applications, including but not limited to teleoperation of anthropomorphic robot hands or surgical robots, virtual and augmented reality, and collection of human motion data.

Therapy effect on hand function after home use of a wearable assistive soft-robotic glove supporting grip strength.

BACKGROUND: Soft-robotic gloves with an assist-as-needed control have the ability to assist daily activities where needed, while stimulating active and highly functional movements within the user's possibilities. Employment of hand activities with glove support might act as training for unsupported hand function. OBJECTIVE: To evaluate the therapeutic effect of a grip-supporting soft-robotic glove as an assistive device at home during daily activities. METHODS: This multicentre intervention trial consisted of 3 pre-assessments (averaged if steady state = PRE), one post-assessment (POST), and one follow-up assessment (FU). Participants with chronic hand function limitations were included. Participants used the Carbonhand glove during six weeks in their home environment on their most affected hand. They were free to choose which activities to use the glove with and for how long. The primary outcome measure was grip strength, secondary outcome measures were pinch strength, hand function and glove use time. RESULTS: 63 patients with limitations in hand function resulting from various disorders were included. Significant improvements (difference PRE-POST) were found for grip strength (+1.9 kg, CI 0.8 to 3.1; p = 0.002) and hand function, as measured by Jebson-Taylor Hand Function Test (-7.7 s, CI -13.4 to -1.9; p = 0.002) and Action Research Arm Test (+1.0 point, IQR 2.0; p≤0.001). Improvements persisted at FU. Pinch strength improved slightly in all fingers over six-week glove use, however these differences didn't achieve significance. Participants used the soft-robotic glove for a total average of 33.0 hours (SD 35.3), equivalent to 330 min/week (SD 354) or 47 min/day (SD 51). No serious adverse events occurred. CONCLUSION: The present findings showed that six weeks use of a grip-supporting soft-robotic glove as an assistive device at home resulted in a therapeutic effect on unsupported grip strength and hand function. The glove use time also showed that this wearable, lightweight glove was able to assist participants with the performance of daily tasks for prolonged periods.

An Overall Automated Architecture Based on the Tapping Test Measurement Protocol: Hand Dexterity Assessment through an Innovative Objective Method.

The present work focuses on the tapping test, which is a method that is commonly used in the literature to assess dexterity, speed, and motor coordination by repeatedly moving fingers, performing a tapping action on a flat surface. During the test, the activation of specific brain regions enhances fine motor abilities, improving motor control. The research also explores neuromuscular and biomechanical factors related to finger dexterity, revealing neuroplastic adaptation to repetitive movements. To give an objective evaluation of all cited physiological aspects, this work proposes a measurement architecture consisting of the following: (i) a novel measurement protocol to assess the coordinative and conditional capabilities of a population of participants; (ii) a suitable measurement platform, consisting of synchronized and non-invasive inertial sensors to be worn at finger level; (iii) a data analysis processing stage, able to provide the final user (medical doctor or training coach) with a plethora of useful information about the carried-out tests, going far beyond state-of-the-art results from classical tapping test examinations. Particularly, the proposed study underscores the importance interdigital autonomy for complex finger motions, despite the challenges posed by anatomical connections; this deepens our understanding of upper limb coordination and the impact of neuroplasticity, holding significance for motor abilities assessment, improvement, and therapeutic strategies to enhance finger precision. The proof-of-concept test is performed by considering a population of college students. The obtained results allow us to consider the proposed architecture to be valuable for many application scenarios, such as the ones related to neurodegenerative disease evolution monitoring.

Comparison of Six Sensor Fusion Algorithms with Electrogoniometer Estimation of Wrist Angle in Simulated Work Tasks.

Hand-intensive work is strongly associated with work-related musculoskeletal disorders (WMSDs) of the hand/wrist and other upper body regions across diverse occupations, including office work, manufacturing, services, and healthcare. Addressing the prevalence of WMSDs requires reliable and practical exposure measurements. Traditional methods like electrogoniometry and optical motion capture, while reliable, are expensive and impractical for field use. In contrast, small inertial measurement units (IMUs) may provide a cost-effective, time-efficient, and user-friendly alternative for measuring hand/wrist posture during real work. This study compared six orientation algorithms for estimating wrist angles with an electrogoniometer, the current gold standard in field settings. Six participants performed five simulated hand-intensive work tasks (involving considerable wrist velocity and/or hand force) and one standardised hand movement. Three multiplicative Kalman filter algorithms with different smoothers and constraints showed the highest agreement with the goniometer. These algorithms exhibited median correlation coefficients of 0.75-0.78 for flexion/extension and 0.64 for radial/ulnar deviation across the six subjects and five tasks. They also ranked in the top three for the lowest mean absolute differences from the goniometer at the 10th, 50th, and 90th percentiles of wrist flexion/extension (9.3°, 2.9°, and 7.4°, respectively). Although the results of this study are not fully acceptable for practical field use, especially for some work tasks, they indicate that IMU-based wrist angle estimation may be useful in occupational risk assessments after further improvements.

A Review of Motor Brain-Computer Interfaces Using Intracranial Electroencephalography Based on Surface Electrodes and Depth Electrodes.

Brain-computer interfaces (BCIs) provide a communication interface between the brain and external devices and have the potential to restore communication and control in patients with neurological injury or disease. For the invasive BCIs, most studies recruited participants from hospitals requiring invasive device implantation. Three widely used clinical invasive devices that have the potential for BCIs applications include surface electrodes used in electrocorticography (ECoG) and depth electrodes used in Stereo-electroencephalography (SEEG) and deep brain stimulation (DBS). This review focused on BCIs research using surface (ECoG) and depth electrodes (including SEEG, and DBS electrodes) for movement decoding on human subjects. Unlike previous reviews, the findings presented here are from the perspective of the decoding target or task. In detail, five tasks will be considered, consisting of the kinematic decoding, kinetic decoding,identification of body parts, dexterous hand decoding, and motion intention decoding. The typical studies are surveyed and analyzed. The reviewed literature demonstrated a distributed motor-related network that spanned multiple brain regions. Comparison between surface and depth studies demonstrated that richer information can be obtained using surface electrodes. With regard to the decoding algorithms, deep learning exhibited superior performance using raw signals than traditional machine learning algorithms. Despite the promising achievement made by the open-loop BCIs, closed-loop BCIs with sensory feedback are still in their early stage, and the chronic implantation of both ECoG surface and depth electrodes has not been thoroughly evaluated.

Word learning is hands-on: Insights from studying natural behavior.

Infants' interactions with social partners are richly multimodal. Dyads respond to and coordinate their visual attention, gestures, vocalizations, speech, manual actions, and manipulations of objects. Although infants are typically described as active learners, previous experimental research has often focused on how infants learn from stimuli that is well-crafted by researchers. Recent research studying naturalistic, free-flowing interactions has explored the meaningful patterns in dyadic behavior that relate to language learning. Infants' manual engagement and exploration of objects supports their visual attention, creates salient and diverse views of objects, and elicits labeling utterances from parents. In this chapter, we discuss how the cascade of behaviors created by infant multimodal attention plays a fundamental role in shaping their learning environment, supporting real-time word learning and predicting later vocabulary size. We draw from recent at-home and cross-cultural research to test the validity of our mechanistic pathway and discuss why hands matter so much for learning. Our goal is to convey the critical need for developmental scientists to study natural behavior and move beyond our "tried-and-true" paradigms, like screen-based tasks. By studying natural behavior, the role of infants' hands in early language learning was revealed-though it was a behavior that was often uncoded, undiscussed, or not even allowed in decades of previous research. When we study infants in their natural environment, they can show us how they learn about and explore their world. Word learning is hands-on.

Dynamic muscular endurance as an indicator of functional readiness of cyber-athletes.

OBJECTIVE: Aim: To study the dynamic muscular endurance of hand movement according to the tapping test in connection with the manifestations of cognitive qualities of cyber-athletes and students involved in computer games as a hobby. PATIENTS AND METHODS: Materials and Methods: Dynamic muscular endurance of the right and left hands of the examined subjects was studied (using the tapping test method), as well as the reaction to a moving object using the diagnostic complex "Diagnost-1". Correction tables (Landolt rings) were used to study voluntary attention. 45 students of the National University of Ukraine on Physical Education and Sport of both sexes, aged 17-26, took part in the study, among whom 10 are cyber-athletes (sports experience of 1-10 years), 15 amateurs (involved in computer games as a hobby) and 20 students who do not play computer games (control group). RESULTS: Results: In cyber-athletes and students involved in computer games, the dynamic muscular endurance of the movement of the hand of the subdominant hand was greater than in students who did not engage in computer games. A higher level of dynamic muscular endurance for the subdominant hand and less functional asymmetry according to the tapping test scores in cyber-athletes were associated with a more successful performance of the attention test. CONCLUSION: Conclusions: A higher level of dynamic muscular endurance for the subdominant arm and a smaller functional asymmetry according to the tapping test indicators in e-athletes can be considered as an indicator of functional readiness.

Scalp acupuncture guidance for identifying the optimal site for transcranial electrical stimulation of the hand.

Transcranial electrical stimulation (tES) often targets the EEG-guided C3/C4 area that may not accurately represent M1 for hand muscles. This study aimed to determine if the neuroanatomy-based scalp acupuncture-guided site (AC) was a more effective spot than the C3 site for neuromodulation. Fifteen healthy subjects received one 20-minute session of high-definition transcranial alternating current stimulation (HD-tACS) intervention (20 Hz at 2 mA) at the AC or C3 sites randomly with a 1-week washout period. Subjects performed ball-squeezing exercises with the dominant hand during the HD-tACS intervention. The AC site was indiscernible from the finger flexor hotspot detected by TMS. At the baseline, the MEP amplitude from finger flexors was greater with less variability at the AC site than at the C3 site. HD-tACS intervention at the AC site significantly increased the MEP amplitude. However, no significant changes were observed after tACS was applied to the C3 site. Our results provide evidence that HD-tACS at the AC site produces better neuromodulation effects on the flexor digitorum superficialis (FDS) muscle compared to the C3 site. The AC localization approach can be used for future tES studies.

The role of different acoustic environmental stimuli on manual dexterity.

Music has been reported to facilitate motor performance. However, there is no data on the effects of different acoustic environmental stimuli on manual dexterity. The present observational study aimed at investigating the effects of background music and noise on a manual dexterity task in young, middle-aged and elderly subjects. Sixty healthy, right-handed subjects aged between 18 and 80 years were enrolled. Twenty young (mean age: 22±2 years), 20 middle-aged (mean age: 55±8 years) and 20 elderly (mean age: 72±5 years) subjects performed the Nine Hole Peg Test (NHPT) in four different acoustic environments: silence (noise < 20dBA), classical music at 60dBA, rock music at 70 dBA, and a noise stimulus at 80dBA. Performance was recorded using an optical motion capture system and retro-reflective markers (SMART DX, 400, BTS). Outcome measures included the total test time and peg-grasp, peg-transfer, peg-in-hole, hand-return, and removing phases times. Normalized jerk, mean and peak of velocity during transfer and return phases were also computed. No differences were found for NHPT phases and total times, normalized jerk, peak of velocity and mean velocity between four acoustic conditions (p>0.05). Between-group differences were found for NHPT total time, where young subjects revealed better performance than elderly (p˂0.001) and middle-aged (p˂0.001) groups. Music and noise stimuli in the considered range of intensity had no influence on the execution of a manual dexterity task in young, middle-aged and elderly subjects. These findings may have implications for working, sportive and rehabilitative activities.

Adaptation of the gain of the corrective lifting response in object manipulation transfers across the hand.

Our ability to skillfully manipulate objects is supported by rapid corrective responses that are initiated when we experience perturbations that interfere with movement goals. For example, the corrective lifting response is triggered when an object is heavier than expected and fails to lift off the surface. In this situation, the absence of expected sensory feedback signalling lift off initiates, within ~ 90 ms, an increase in lifting force. Importantly, when people repeatedly lift an object that, on occasional catch trials, is heavier than expected, the gain of the corrective response, defined as the rate of force increase, adapts to the 'catch' weight. In the present study, we investigated whether this response adaption transfers intermanually. In the training phase, participants used either their left or right hand (counterbalanced) to repeatedly lift a 3 N object that unexpectedly increased to 9 N on catch trials, leading to an increase in the gain of the lifting response across catch trials. Participants then lifted the object with their other hand. On the first catch trial, the gain remained elevated and thus transferred across the hands. This finding suggests that the history of lifts performed by one hand updates the corrective responses for both hands.

Identifying alterations in hand movement coordination from chronic stroke survivors using a wearable high-density EMG sleeve.

Objective.Non-invasive, high-density electromyography (HD-EMG) has emerged as a useful tool to collect a range of neurophysiological motor information. Recent studies have demonstrated changes in EMG features that occur after stroke, which correlate with functional ability, highlighting their potential use as biomarkers. However, previous studies have largely explored these EMG features in isolation with individual electrodes to assess gross movements, limiting their potential clinical utility. This study aims to predict hand function of stroke survivors by combining interpretable features extracted from a wearable HD-EMG forearm sleeve.Approach.Here, able-bodied (N= 7) and chronic stroke subjects (N= 7) performed 12 functional hand and wrist movements while HD-EMG was recorded using a wearable sleeve. A variety of HD-EMG features, or views, were decomposed to assess alterations in motor coordination.Main Results.Stroke subjects, on average, had higher co-contraction and reduced muscle coupling when attempting to open their hand and actuate their thumb. Additionally, muscle synergies decomposed in the stroke population were relatively preserved, with a large spatial overlap in composition of matched synergies. Alterations in synergy composition demonstrated reduced coupling between digit extensors and muscles that actuate the thumb, as well as an increase in flexor activity in the stroke group. Average synergy activations during movements revealed differences in coordination, highlighting overactivation of antagonist muscles and compensatory strategies. When combining co-contraction and muscle synergy features, the first principal component was strongly correlated with upper-extremity Fugl Meyer hand sub-score of stroke participants (R2= 0.86). Principal component embeddings of individual features revealed interpretable measures of motor coordination and muscle coupling alterations.Significance.These results demonstrate the feasibility of predicting motor function through features decomposed from a wearable HD-EMG sleeve, which could be leveraged to improve stroke research and clinical care.

Pre-movement muscle co-contraction associated with motor performance deterioration under high reward conditions.

Reward usually enhances task performance, but exceptionally large rewards can impede performance due to psychological pressure. In this study, we investigated motor activity changes in high-reward situations and identified indicators for performance decline. Fourteen healthy adults practiced a velocity-dependent right-hand motor task for three days, followed by a test day with varying monetary reward for each trial. Participants were divided into low performers (LPs) and high performers (HPs) according to whether success rate decreased or increased, respectively, on the highest reward trials compared to lower reward trials. Both LPs and HPs demonstrated increased hand velocity during higher reward trials, but only LPs exhibited a significant increase in velocity variance. There was also a negative correlation between the pre-movement co-contraction index (CCI) of the biceps and triceps muscles and success rate on the highest reward trials. This correlation was confirmed in a second experiment with 12 newly recruited participants, suggesting that pre-movement CCI is a marker for performance decline caused by high reward. These findings suggest that interventions to reduce pre-movement CCI such as biofeedback training could be useful for preventing the paradoxical decline in motor performance associated with high rewards.

Pink noise promotes sooner state transitions during bimanual coordination.

The seemingly straightforward task of tying one's shoes requires a sophisticated interplay of joints, muscles, and neural pathways, posing a formidable challenge for researchers studying the intricacies of coordination. A widely accepted framework for measuring coordinated behavior is the Haken-Kelso-Bunz (HKB) model. However, a significant limitation of this model is its lack of accounting for the diverse variability structures inherent in the coordinated systems it frequently models. Variability is a pervasive phenomenon across various biological and physical systems, and it changes in healthy adults, older adults, and pathological populations. Here, we show, both empirically and with simulations, that manipulating the variability in coordinated movements significantly impacts the ability to change coordination patterns-a fundamental feature of the HKB model. Our results demonstrate that synchronized bimanual coordination, mirroring a state of healthy variability, instigates earlier transitions of coordinated movements compared to other variability conditions. This suggests a heightened adaptability when movements possess a healthy variability. We anticipate our study to show the necessity of adapting the HKB model to encompass variability, particularly in predictive applications such as neuroimaging, cognition, skill development, biomechanics, and beyond.

A tactile sensing system capable of recognizing objects based on bioinspired self-sensing soft pneumatic actuator.

Tactile sensors play an important role when robots perform contact tasks, such as physical information collection, force or displacement control to avoid collision. For these manipulations, excessive contact may cause damage while poor contact cause information loss between the robotic end-effector and the objects. Inspired by skin structure and signal transmission method, this paper proposes a tactile sensing system based on the self-sensing soft pneumatic actuator (S-SPA) capable of providing tactile sensing capability for robots. Based on the adjustable height and compliance characteristics of the S-SPA, the contact process is safe and more tactile information can be collected. And to demonstrate the feasibility and advantage of this system, a robotic hand with S-SPAs could recognize different textures and stiffness of the objects by touching and pinching behaviours to collect physical information of the various objects under the positive work states of the S-SPA. The result shows the recognition accuracy of the fifteen texture plates reaches 99.4%, and the recognition accuracy of the four stiffness cuboids reaches 100%by training a KNN model. This safe and simple tactile sensing system with high recognition accuracies based on S-SPA shows great potential in robotic manipulations and is beneficial to applications in domestic and industrial fields.

Distribution of control during bimanual movement and stabilization.

In two-handed actions like baseball batting, the brain can allocate the control to each arm in an infinite number of ways. According to hemispheric specialization theory, the dominant hemisphere is adept at ballistic control, while the non-dominant hemisphere is specialized at postural stabilization, so the brain should divide the control between the arms according to their respective specialization. Here, we tested this prediction by examining how the brain shares the control between the dominant and non-dominant arms during bimanual reaching and postural stabilization. Participants reached with both hands, which were tied together by a stiff virtual spring, to a target surrounded by an unstable repulsive force field. If the brain exploits each hemisphere's specialization, then the dominant arm should be responsible for acceleration early in the movement, and the non-dominant arm will be the prime actor at the end when holding steady against the force field. The power grasp force, which signifies the postural stability of each arm, peaked at movement termination but was equally large in both arms. Furthermore, the brain predominantly used the arm that could use the stronger flexor muscles to mainly accelerate the movement. These results point to the brain flexibly allocating the control to each arm according to the task goal without adhering to a strict specialization scheme.

Dual Stream Long Short-Term Memory Feature Fusion Classifier for Surface Electromyography Gesture Recognition.

Gesture recognition using electromyography (EMG) signals has prevailed recently in the field of human-computer interactions for controlling intelligent prosthetics. Currently, machine learning and deep learning are the two most commonly employed methods for classifying hand gestures. Despite traditional machine learning methods already achieving impressive performance, it is still a huge amount of work to carry out feature extraction manually. The existing deep learning methods utilize complex neural network architectures to achieve higher accuracy, which will suffer from overfitting, insufficient adaptability, and low recognition accuracy. To improve the existing phenomenon, a novel lightweight model named dual stream LSTM feature fusion classifier is proposed based on the concatenation of five time-domain features of EMG signals and raw data, which are both processed with one-dimensional convolutional neural networks and LSTM layers to carry out the classification. The proposed method can effectively capture global features of EMG signals using a simple architecture, which means less computational cost. An experiment is conducted on a public DB1 dataset with 52 gestures, and each of the 27 subjects repeats every gesture 10 times. The accuracy rate achieved by the model is 89.66%, which is comparable to that achieved by more complex deep learning neural networks, and the inference time for each gesture is 87.6 ms, which can also be implied in a real-time control system. The proposed model is validated using a subject-wise experiment on 10 out of the 40 subjects in the DB2 dataset, achieving a mean accuracy of 91.74%. This is illustrated by its ability to fuse time-domain features and raw data to extract more effective information from the sEMG signal and select an appropriate, efficient, lightweight network to enhance the recognition results.