Electromyography Sensors Research Articles

SEMG-Based hand gestures classification using a semi-supervised multi-layer neural networks with Autoencoder

This work presents a semi-supervised multilayer neural network (MLNN) with an Autoencoder to develop a classification model for recognizing hand gestures from electromyographic (EMG) signals. Using a Myo armband equipped with eight non-invasive surface-mounted biosensors, raw surface EMG (sEMG) sensor data were captured corresponding to five hand gestures: Fist, Open hand, Wave in, Wave out, and Double tap. The sensor collected data underwent preprocessing, feature extraction, label assignment, and dataset organization for classification tasks. The model implementation, validation, and testing demonstrated its efficacy after incorporating synthetic sEMG data generated by an Autoencoder. In comparison to the state-of-the-art techniques from the literature, the proposed model exhibited strong performance, achieving accuracy of 99.68%, 100%, and 99.26% during training, validation, and testing, respectively. Comparatively, the proposed MLNN with Autoencoder model outperformed a K-Nearest Neighbors model established for comparative evaluation.

Multichannel EMG-based gesture recognition utilizing advanced machine learning techniques: A random forest classifier for high-precision signal classification

This research examines how advanced machine learning algorithms can be used to classify multichannel electromyographic (EMG) signals with a high level of accuracy to assist in recognizing hand gestures. The goal is to create a robust and scalable system for gesture-based virtual control using EMG signals with potential applications in assistive technologies, rehabilitation, and human-computer interaction. Data were gathered using a MYO Thalmic bracelet containing eight EMG sensors on thirty-six subjects, and a Random Forest classifier was trained to identify seven distinct types of hand gestures (rest, fist clench, wrist flexion/extension, and radial/ulnar deviations). The machine learning pipeline included extensive preprocessing (i.e., EMG signal normalization and signal feature extraction; root mean square, waveform length, and zero crossing rate) and several hyperparameter tuning procedures to improve model performance. The Random Forest model (100 decision trees) achieved an overall classification accuracy of 98.68%, with a range of accuracies for each class (e.g., 95.2% wrist flexion and 91.8% ulnar deviation) when evaluated using cross-validation (i.e., average F1-score = 0.92, precision = 0.94, recall = .91). Overall, the study provides strong evidence for the effectiveness of ensemble learning methods at analyzing complex, multidimensional EMG signals. The high classification accuracy reported, in particular, demonstrates that the system could function for real-time recognition of hand gestures in a virtual environment. Ultimately, the initial work sets the stage for future exploration of a model that may be integrated with actuation models to control prosthetic limbs, virtual actors/avatars, and robotic devices. By demonstrating a scalable and efficient method of gesture recognition using EMG signals, these early findings enable future pathways and possibilities to design innovative, assistive solutions for digital systems that increase accessibility and interaction for users who are motor impaired or have a limited range of motion.

Biomechanical Analysis of Injury Risk in Two High-Altitude Landing Positions Using Xsens Inertial Units and EMG Sensors.

High-altitude landing maneuvers can pose a significant injury risk, particularly when performed with different landing techniques. This study aims to compare the biomechanical parameters and injury risks associated with two landing positions-staggered foot landing and simultaneous bilateral landing-using Xsens inertial units and electromyography (EMG) sensors. A total of 26 university students (13 males, 13 females) participated in this study. Kinematic data were collected using inertial measurement units (IMUs), muscle activity was recorded with EMG, and ground reaction forces were captured using 3D force plates. The data were processed and analyzed using the AnyBody modeling system to simulate joint forces, moments, and muscle activation. This study found that simultaneous bilateral landing exhibited greater hip flexion-extension, knee flexion-extension, and ankle inversion. Vertical joint forces were also significantly higher in the hip, knee, and ankle during simultaneous bilateral landing. Staggered foot landing showed higher muscle forces in the gluteus maximus, iliopsoas, and quadriceps femoris (p < 0.001). The EMG analysis revealed significant differences in the biceps femoris (p = 0.008) and quadriceps femoris (p < 0.001). These findings suggest that simultaneous bilateral landing increases joint load, while staggered foot landing increases muscle activation, which may lead to different injury risks between the two techniques.

Three-dimensional finite element analysis of biological signal feedback in the mechanical properties of sound production

In response to the problems of low reliability and long time consumption in traditional research on sound production, this article used electromyography (EMG) signals as input signals to build a high-precision sound production mechanics model. A Proportional-Integral-Derivative (PID) controller was used to dynamically adjust the model and develop a real-time feedback system. This article established a detailed three-dimensional (3D) finite element model including the vocal cords and throat, defined the nonlinear elastic properties and linear elastic properties of different tissues, and used tetrahedral grid partitioning technology to improve the computational accuracy of the model. Through a EMG sensor, an individual EMG was collected and filtered to remove noise. The processed EMGs were used as input parameters for the finite element model to drive the muscle units in the model. By using a PID controller to receive real-time EMG input, the error was calculated and the model was adjusted, enabling accurate simulation of the mechanical properties of vocal cord vibration under different vocal states and achieving real-time feedback. Considering the complexity of vocal cord vibration driven by biological signals, this article simulated and analyzed the modal characteristics of vocal cord vibration, and analyzed the differences in vocal cord vibration characteristics under different vocal states. The sound pressure distribution and resonance frequency were simulated and analyzed to understand the propagation characteristics of sound. Finally, by comparing and analyzing the simulated data with the actual collected data, it was found that the maximum relative error rate of the model under different sound states was 6.14%, and the overall error rate of the model was relatively low, which verified the reliability of the model. The findings demonstrated that the feedback delays of the model in different sound states were all within 100 milliseconds, indicating that the system had high real-time performance and accuracy, which was promising in practical applications.

Comparison of Lower-Limb Muscle Synergies Between Young and Old People During Cycling Based on Electromyography Sensors—A Preliminary Cross-Sectional Study

The purpose of this study was to analyze the lower-limb muscle synergies of young and older adults during stationary cycling across various mechanical conditions to reveal adaptive strategies employed by the elderly to address various common pedaling tasks and function degradation. By comparing lower-limb muscle synergies during stationary cycling between young and old people, this study examined changes in muscle synergy patterns during exercise in older individuals. This is crucial for understanding neuromuscular degeneration and changes in movement patterns in older individuals. Sixteen young and sixteen older experienced cyclists were recruited to perform stationary cycling tasks at two levels of power (60 and 100 W) and three cadences (40, 60, and 90 rpm) in random order. The lower-limb muscle synergies and their inter- and intra-individual variability were analyzed. Three synergies were extracted in this study under all riding conditions in both groups while satisfying overall variance accounted for (VAF) &gt; 85% and muscle VAF &gt; 75%. The older adults exhibited lower variability in synergy vector two and a higher trend in the variability of activation coefficient three, as determined by calculating the variance ratio. Further analyses of muscle synergy structures revealed increased weighting in major contribution muscles, the forward-shifting peak activation in synergy one, and lower peak magnitude in synergy three among older adults. To produce the same cycling power and cadence as younger individuals, older adults make adaptive adjustments in muscle control—increased weighting in major contribution muscles, greater consistency in the use of primary force-producing synergies, and earlier peak activation of subsequent synergy.

Pattern Recognition in the Processing of Electromyographic Signals for Selected Expressions of Polish Sign Language.

Gesture recognition has become a significant part of human-machine interaction, particularly when verbal interaction is not feasible. The rapid development of biomedical sensing and machine learning algorithms, including electromyography (EMG) and convolutional neural networks (CNNs), has enabled the interpretation of sign languages, including the Polish Sign Language, based on EMG signals. The objective was to classify the game control gestures and Polish Sign Language gestures recorded specifically for this study using two different data acquisition systems: BIOPAC MP36 and MyoWare 2.0. We compared the classification performance of various machine learning algorithms, with a particular emphasis on CNNs on the dataset of EMG signals representing 24 gestures, recorded using both types of EMG sensors. The results (98.324% versus ≤7.8571% and 95.5307% versus ≤10.2697% of accuracy for CNNs and other classifiers in data recorded with BIOPAC MP36 and MyoWare, respectively) indicate that CNNs demonstrate superior accuracy. These results suggest the feasibility of using lower-cost sensors for effective gesture classification and the viability of integrating affordable EMG-based technologies into broader gesture recognition frameworks, providing a cost-effective solution for real-world applications. The dataset created during the study offers a basis for future studies on EMG-based recognition of Polish Sign Language.

A Multimodal Bracelet to Acquire Muscular Activity and Gyroscopic Data to Study Sensor Fusion for Intent Detection.

Researchers have attempted to control robotic hands and prostheses through biosignals but could not match the human hand. Surface electromyography records electrical muscle activity using non-invasive electrodes and has been the primary method in most studies. While surface electromyography-based hand motion decoding shows promise, it has not yet met the requirements for reliable use. Combining different sensing modalities has been shown to improve hand gesture classification accuracy. This work introduces a multimodal bracelet that integrates a 24-channel force myography system with six commercial surface electromyography sensors, each containing a six-axis inertial measurement unit. The device's functionality was tested by acquiring muscular activity with the proposed device from five participants performing five different gestures in a random order. A random forest model was then used to classify the performed gestures from the acquired signal. The results confirmed the device's functionality, making it suitable to study sensor fusion for intent detection in future studies. The results showed that combining all modalities yielded the highest classification accuracies across all participants, reaching 92.3±2.6% on average, effectively reducing misclassifications by 37% and 22% compared to using surface electromyography and force myography individually as input signals, respectively. This demonstrates the potential benefits of sensor fusion for more robust and accurate hand gesture classification and paves the way for advanced control of robotic and prosthetic hands.

Modeling Muscle Activities of Squat Motion using OpenSIM

Human beings need to be able to monitor their muscle health. Muscle health can be assessed by directly measuring its signal using an electromyography sensor or, recently, from a musculoskeletal modeling software such as OpenSIM. OpenSIM is an open-source platform for modeling, simulating, and analyzing the neuromusculoskeletal system. This work aims to simulate muscle forces during squat movement using OpenSIM simulation and validate the results using electromyography sensors attached to the body. The squat movement data generated from the optical motion capture system and the force plate will be used as inputs for the OpenSIM software. OpenSIM can provide output, such as a graph of the working muscle forces. That graph will be compared and analyzed with a surface electromyography sensor results graph. Based on the analysis results, the chart of working muscle forces generated by the OpenSIM software has shown a similar trend to the graph of muscle activity resulting from surface electromyography sensor reading. Therefore, the simulation results from the OpenSIM software were validated, and it was implied that the data collection process and the modeling were done correctly.

Digital Twin for Amputees: A Bidirectional Interaction Modeling and Prototype with Convolutional Neural Network

With over 2 million amputees in the U.S., they have been facing significant employment challenges. However, several physical prostheses still lack usability to be able for them to work in person. Therefore, this study explores an innovative application of digital twin approach, focusing on bidirectional interaction modeling and prototyping using convolutional neural networks (CNNs). We developed a simplified digital twin environment integrating electromyography (EMG) sensors and virtual reality (VR) to enable real-time interaction between the virtual and physical worlds. The CNN model, trained to classify hand movement from EMG data, achieved a test accuracy of 99%, demonstrating its effectiveness for practical applications. Our framework facilitates remote control of physical devices through VR gestures, potentially allowing amputees to perform meaningful work from home, thus overcoming physical limitations and fostering greater independence. This preliminary study underscores the possibility of digital twin technology to redefine workplace accessibility, offering amputees opportunities for potential employment.

Hypospadias Reconstruction Training: Development of an Ex-Vivo Model for Objective Evaluation of Surgical Skills

To objectively evaluate technical skill acquisition in hypospadias repair procedures during surgical training using noninvasive wearable sensor technology. We combined subjective video evaluations with objective electromyography (EMG) measurements in a hands-on hypospadias training course. Surgeons wore wireless EMG and accelerometer sensors on their dominant hand while performing tasks on ex-vivo cadaveric calf penises. The study focused on 4 skills as follows: urethral mobilization, dorsal inlay graft harvest/implantation, meatal-based flap urethroplasty, and dorsal plication. Machine learning techniques analyzed muscle activation patterns and attributes for assessing surgical precision. The course included 18 participants (10 female, 8 males; average age 40.18±8.46 years) categorized as novice (n=10, <3 years' experience), intermediate (n=5, 3-5 years), and expert (n=3, >5 years). Video evaluations did not reveal significant differences due to short-term training. However, EMG measurements showed significant reductions in average EMG power, total time, dominant frequency, and cumulative muscle workload after training. Additionally, the mean power spectral density of the EMG signal decreased notably post-training. This study presents a structured approach for hypospadias training and highlights the effectiveness of wearable sensor technology for objective skill assessment. While video evaluations did not detect significant changes, EMG data provided measurable differences in skill acquisition, suggesting that wearable sensors could enhance objective evaluations of surgical proficiency in residency programs.

Commercial infant products influence body position and muscle use

The musculoskeletal and motor development of infants is affected by their environment, which varies from being held, lying on a firm flat surface, to seated in various nursery products. Nursery products can alter the body position of infants, particularly the position of the head/neck and trunk, which may inhibit an infant's ability to breathe. With U.S. infants spending an increasing amount of time in seated products, the purpose of this study was to assess muscle activation and body position in four commercial infant products (carrier, bouncer, rocker, and swing) during supine and prone positioning, compared to a firm flat surface. Thirteen healthy infants (age: 4.2 ± 1.4 months; 7 M/6F) were enrolled in this IRB-approved in-vivo biomechanics study. Surface electromyography sensors recorded muscle activity of the erector spinae, cervical paraspinals, quadriceps, and abdominal muscles and retro-reflective markers tracked movements to determine head-neck, trunk, and torso-pelvis flexion/extension in the sagittal plane. While supine, infants exhibited increased head-neck and trunk flexion of up to 21° and 27° above the playmat, respectively, in all seated products. While prone, high abdominal muscle activation compared to the playmat indicates that infants will fatigue faster in seated products. Additionally, the lower muscle activation levels exhibited in younger infants (< 4 months) compared to older infants (≥ 4 months) indicates that younger infants rely on the product design to maintain body position. However, offering infants a variety of environments to move within is important to avoid motor delay, therefore future work should explore how long-term use may impact an infant's development.

Impact of novel lightweight disposable gastroscope and duodenoscope on endoscopist muscle activation: a comparative study

Background and AimsMusculoskeletal injuries are common among endoscopists and reducing the endoscope weight is considered a preventive measure. We aimed to assess the impact on the muscle activation of recently developed lightweight gastroscopes and duodenoscopes, compared with standard endoscopes in a two-part study. MethodsFirstly, fourteen participants performed a protocol using a lightweight disposable gastroscope or a standard reusable in a random order. The protocol constituted of three equal working blocks with repetitive and standardized movements. In the second part, fifteen endoscopists used a lightweight disposable duodenoscope and a reusable in an artificial model and at rest. For both protocols, the subjects wore a sleeve with embedded superficial electromyographic sensors on the left arm. ResultsWrist flexor muscle activation was significantly lower when using the lightweight single-use endoscopes during the respective protocol tasks (P<0.05). The extensor muscles were also significantly less activated while using the single-use gastroscope (P<0.05), but there was no difference when using the duodenoscopes. ConclusionsStandardized operation with lightweight endoscopes reduces the load on the left forearm muscles and favors muscle rest compared to the heavier standard endoscopes. These improvements in user ergonomics may aid in preventing or delaying the development of tremor, fatigue, and injuries.

The Influence of Body Position on the Resting Motor Threshold of Posterior Root-Muscle Reflexes Evoked via Transcutaneous Spinal Cord Stimulation.

Background: Thoracolumbar transcutaneous spinal cord stimulation (tSCS) non-invasively evokes posterior root-muscle reflexes (PRMR) with the aim of neuromodulating sensorimotor function following spinal cord injury. Research is still in its infancy regarding the effect of body position on the nature of these spinally evoked responses. Therefore, the aim of this study was to investigate the influence of body position on the nature of PRMR responses during tSCS. Methods: A total of 11 (6M, 5F) participants completed a full PRMR recruitment curve from 10 ma up to 120 ma (10 ma increments) at the T11/12 intervertebral space using a singular 3.2 cm diameter cathode. At each intensity, three paired pulses (50 ms inter-pulse interval), followed by three singular pulses with a six-second delay were applied in each body position (supine, supine 90-90, sitting and standing) in a randomised order. The PRMR responses in lower limb muscles were recorded using wireless electromyographic sensors placed on the Soleus, Tibialis Anterior, Rectus Femoris and Bicep Femoris long head. A two-way (body position × muscle) repeated measures analysis of variance was used to investigate the effect of body position on PRMR-evoked responses. Results: There was a significant main effect of body position on PRMR resting motor threshold (RMT) (p < 0.001), first response peak-to-peak amplitude (p = 0.003) and percentage post-activation depression (%PAD) (p = 0.012). Sitting had significantly higher RMT and significantly lower first response peak-to-peak amplitudes compared to all other positions, but significant differences in %PAD were only detectible between supine and standing. Conclusions: Body position influences the nature of PRMR-evoked responses during tSCS.

Predicting Perceived Back Fatigue During Exoskeleton Supported Trunk Bending Tasks using Machine Learning

Repetitive trunk flexion tasks performed over long durations can increase low-back injury risk, where Back Support Industrial Exoskeletons (BSIEs) can be beneficial. While BSIEs have shown effectiveness in lab assessments, real-world outcomes have shown variation based on task complexity, necessitating monitoring of physical demands. Fourteen participants performed repetitive trunk BSIE-assisted forward bending and return, without fatigue and then at medium-high fatigue. We recorded muscle activity in low-back and thigh muscles using Electromyography (EMG) and whole-body stability using force plates. Classification algorithms, namely, Support Vector Machine (SVM), Random Forest (RF), and XGBoost (XGB) were utilized to predict perceived medium-high back fatigue using sensor data. Highest performance was observed with XGB algorithm using data from a single low-back EMG sensor (Accuracy: 86.1%, Recall: 86%), and force plate (93.5, 94.1%). Outcomes of our study can be helpful in developing novel fatigue detection products, benefiting ergonomists in properly implementing BSIEs in industrial scenarios.

Comparative Electromyographic Activity of Hamstrings During Sprinting Versus Strengthening Exercises: Implications for Injury Prevention.

Hamstring injuries are a major problem in sports involving sprinting, such as soccer, rugby, and track and field, and lead to sports stoppages and psychological, social, and financial repercussions. For several years now, these injuries have been stagnating or even increasing. Preventing these injuries is therefore a fundamental issue for at-risk athletes. The aim of this study was to compare the electromyographic(EMG) activity of the hamstrings in athletes during sprinting, Nordic hamstring exercise (NHE), and high-speed concentric exercise on an isokinetic dynamometer. A pilot study was conducted on a population of 15 sprint-exposed field athletes (22.54 ± 3.71 years, Tegner score ≥ 6) with no history of hamstring injury in the last six months. The protocol included a warm-up, followed by three repetitions of the NHE, two sets of 10 repetitions on the isokinetic dynamometer at 300°/sec, and a maximal sprint. Exercises were randomized for each subject, and muscle activity was assessed using wireless EMG sensors during each test. EMG data were normalized to the measured maximum voluntary isometric contraction (MVIC), and test results were statistically analyzed to determine which exercise resulted in maximal hamstring activation. Comparison of hamstring muscle activity between exercises showed a significant difference for most of our results (p-value < 0.05). The results show significantly higher mean hamstring activity during sprinting (0.4800 ± 0.19 μV) compared with strengthening exercises. The NHE (0.3201 ± 0.09 μV) was the second most active exercise. In the last place was the high-speed concentric exercise on the isokinetic machine, which produced less activation than the other exercises (0.2487 ± 0.07 μV). Sports involving sprinting are at risk of a hamstring injury but it appears that its use in rehabilitation and prevention of hamstring injury is relevant, as it would allow high-intensity muscle activation to prepare the hamstring for this type of loading. However, it is also fundamental to integrate strengthening exercises such as NHE in combination with sprinting in our rehabilitation. Finally, the use of the isokinetic dynamometer does not constitute a first-line choice for hamstring injury management.

Optimizing Sensor Placement and Machine Learning Techniques for Accurate Hand Gesture Classification

Millions of individuals are living with upper extremity amputations, making them potential beneficiaries of hand and arm prostheses. While myoelectric prostheses have evolved to meet amputees’ needs, challenges remain related to their control. This research leverages surface electromyography sensors and machine learning techniques to classify five fundamental hand gestures. By utilizing features extracted from electromyography data, we employed a nonlinear, multiple-kernel learning-based support vector machine classifier for gesture recognition. Our dataset encompassed eight young nondisabled participants. Additionally, our study conducted a comparative analysis of five distinct sensor placement configurations. These configurations capture electromyography data associated with index finger and thumb movements, as well as index finger and ring finger movements. We also compared four different classifiers to determine the most capable one to classify hand gestures. The dual-sensor setup strategically placed to capture thumb and index finger movements was the most effective—this dual-sensor setup achieved 90% accuracy for classifying all five gestures using the support vector machine classifier. Furthermore, the application of multiple-kernel learning within the support vector machine classifier showcases its efficacy, achieving the highest classification accuracy amongst all classifiers. This study showcased the potential of surface electromyography sensors and machine learning in enhancing the control and functionality of myoelectric prostheses for individuals with upper extremity amputations.

Ability of Nicotinamide Riboside to Prevent Muscle Fatigue of Barrows Subjected to a Performance Test.

The objective of this study was to determine the daily dietary nicotinamide riboside (NR) dose required to maximize the delay of subjective muscle fatigue onset. Barrows (N = 100) were assigned to one of five treatments: a conventional swine finishing diet containing 0 (CON), 15 (15NR), 30 (30NR), 45 (45NR) mg·kg body weight-1·d-1 NR, or CON supplemented with 45 mg·kg body weight-1·d-1 NR by drench or cookie dough (DRE). All treatments were administered for the final 11 days of feeding. On supplementation d 10, barrows individually experienced a performance test at 1.09 m/s until they were subjectively exhausted. Wireless electromyography (EMG) sensors were affixed to the biceps femoris (BF), tensor fascia latae (TFL), and semitendinosus (ST) to measure real-time muscle activity. There were no treatment effects for barrow speed (p = 0.57), a tendency for a treatment effect (p = 0.07) for distance, and a treatment effect (p = 0.04) on time to exhaustion. Barrows of the 15NR and DRE treatments had greater (p = 0.05) distances to exhaustion than CON barrows but did not differ from other NR barrows (p > 0.11). Barrows in the 45NR treatment did not differ (p = 0.11) in distance from 30NR barrows but tended to have a greater (p = 0.07) distance compared to CON barrows. All other treatment comparisons did not differ (p > 0.27). Barrows in the DRE treatment moved for longer (p < 0.01) than CON barrows, but all other treatments did not differ from each other (p > 0.15). There was no treatment × period interaction for all muscles' root mean square (RMS) values (p > 0.16), but there were Period effects for all muscles (p < 0.01) and a Treatment effect (p = 0.04) in the TFL. For all muscles, period 4 had greater RMS values than all other periods (p < 0.01), who did not differ from each other (p > 0.29). In the TFL, CON barrows had greater RMS values during the performance test compared to all NR treatments (p < 0.02), who did not differ from each other (p > 0.18). Overall, NR demonstrates potential in being a useful tool in fatigue prevention, but efficient administration of the compound needs further investigation.

Development of finger movement assistive gloves with pneumatic fabric actuators

We developed flexible, lightweight, and washable gloves with actuators to assist finger movements and improve ease of wearing. Performance and wearability were measured using standardized tests, triangulation of bending angles, electromyography (EMG), and grip strength. User satisfaction was measured using a survey. EMG sensors were attached to the flexor digitorum superficialis and extensor digitorum communis to capture movement data for grasping and releasing, lifting and putting down, and opening and closing an object with (a) gloves and an actuator, (b) gloves and no actuator, and (c) no gloves. The actuator-equipped glove weighed 31.4 g—lighter than in any earlier studies. In situation (a), the average EMG values for the four participants decreased, ranging from −2.06% to −44.1%, confirming the superior performance of the gloves. Survey results revealed high levels of user satisfaction. Our study offers insights into the development of rehabilitation robotic gloves that assist muscle movements and are easy to wear.

378 Workload Assessment of Suturing Tasks and AI-Task Classification Using Electromyography

Abstract Aim It is well established that excess surgeon workload results in poorer performance. Workload can be divided into physical workload (PWL) and cognitive workload (CWL). The study aims are twofold: 1. Assess how the subjective PWL of suturing is affected by varying PWL and CWL 2. Develop an AI-classification algorithm for task detection using electromyography (EMG). Method Novices were trained in interrupted suturing and asked to suture under four conditions of varied workload. Tasks included were: 1. Control, 2. Different suture material (Silk), 3. Simultaneously undertaking a neurocognitive task and 4. Suturing at depth. Subjective workload was recorded using a validated workload questionnaire (SURG-TLX). EMG sensors recorded physiological data from the biceps brachii and extensor ulnar carpi. A convolutional neural network (CNN) was created for AI-task classification using EMG data. Results The depth task had the highest subjective workload. Subjective depth raw workload was significantly higher than Control (p=&amp;lt;0.01) and Silk (p=&amp;lt;0.001) but comparable to the neurocognitive task. Subjective depth weighted workload was significantly higher than the neurocognitive task (p=&amp;lt;0.05). Subjective depth adjusted workload was significantly higher than the neurocognitive and silk tasks (both p=&amp;lt;0.05). The CNN training accuracy was 76.4% (95%CI 75.5-77.4) and had a classification accuracy of 76.5% (95%CI 75.7-77.2). Conclusions The relationship between PWL and CWL on subjective PWL is complex. Despite this, the CNN was able to differentiate between tasks using objective EMG data with good accuracy. This CNN has potential to characterise CWL which can be further developed to detect excess workload to improve surgical training through objective benchmarks.

Pool-Based Surfboard Elicits Activation of Posterior Shoulder Muscles During a Surfing Stroke.

Pexa, BS, Johnston, CD, Elder, EE, Ford, KR, Patterson, MQ, and Myers, JB. Pool-based surfboard elicits activation of posterior shoulder muscles during a surfing stroke. J Strength Cond Res 38(7): 1300-1304, 2024-Surfboard paddling may activate posterior shoulder muscles, which are critical to baseball pitchers' injury risk and performance. The purpose of this study was to measure posterior shoulder muscle activation during different phases of the surf stroke (propulsion vs. recovery) on a pool-based surfboard. Twenty healthy active adult subjects completed a familiarization and testing session with the pool-based surfboard. During the testing session, electromyography (EMG) sensors were placed on 6 posterior shoulder muscles: latissimus dorsi, infraspinatus, posterior deltoid, upper trapezius, middle trapezius, and lower trapezius. Subjects completed 4 laps in a pool at 3 separate resistances (low, moderate, and heavy) in a randomized order. The peak EMG signal during each phase (propulsion and recovery) was recorded. A 2-way within subject ANOVA (resistance-by-phase) with post hoc Bonferroni's corrections was used to identify differences in EMG activation. There was a significant main effect of phase for the latissimus dorsi (F = 91.3, p < 0.001), upper trapezius (F = 36.5, p < 0.001), middle trapezius (F = 33.8, p < 0.001), and lower trapezius (F = 21.6, p < 0.001). The latissimus dorsi demonstrated higher activation during the propulsion phase (p < 0.001), and all trapezius muscles demonstrated higher activation during the recovery phase (p < 0.001). There was a significant main effect of resistance for the posterior deltoid (F = 3.4, p = 0.043), with higher muscle activation in the low resistance trials compared with the heavy resistance trials (p = 0.036). Recreationally active individuals demonstrate activation of the posterior shoulder when using a pool-based surfboard. This pool-based surfboard may be beneficial to activate the posterior musculature and may be more accessible than standard surfing to baseball athletes.