Explore the vast range of titles available.

Long short-term memory (LSTM) networks are widely used in biomechanical data analysis but have the significant limitations in interpretability and decision transparency. Combining graph neural networks (GNN) with gate recurrent units (GRU) may offer a better solution. This study proposes and validates a hybrid GNN-GRU model for predicting baseball pitching speed and enhancing its interpretability using layer-wise relevance propagation (LRP). C3D data from 53 baseball athletes were downloaded from a public dataset. Kinematic features of 9 joints and pitching speed during the pitching process were calculated using Visual3D, resulting in a total of 208 valid pitches. The feature data were input into both LSTM and GNN-GRU hybrid models, with hyperparameters tuned using particle swarm optimization. LRP was employed to obtain the contribution rate changes of kinematic features to the prediction results throughout the pitching cycle. The prediction accuracy of the models was evaluated using mean absolute error (MAE), mean squared error (MSE), and R-squared (R 2 ). The results showed that there were the significant statistical differences in the MAE and R 2 metrics between the LSTM model and the GNN-GRU model in predicting pitching speed on the test set. The MAE ( P  = 0.000, Z = − 5.170, Cohen’s d  = 1.514) and R 2 ( P  = 0.000, Z = − 2.981, Cohen’s d  = 2.314) of the LSTM model were significantly lower than those of the GNN-GRU model. Compared to LSTM, the GNN-GRU model achieved better prediction accuracy but was potentially more susceptible to the influence of data variability. Moreover, the GNN-GRU-based model demonstrated the better interpretability and decision transparency.

Accurately throwing an object to a target position repeatedly is one of the specific human motor skills. The final arrival position of a thrown ball can be determined by its physical state at release. In baseball pitching, reducing the variability of the velocity angle of the ball at release (release angle) is important for reducing the variability of the pitch location. Although previous studies have suggested that hand and finger movements are important for accurate throwing, their relationship with the release angle has not yet been investigated in detail. This study focused on the positional relationship between the ball and fingers, which is considered to be closely related to ball movement as an action point of the force, and examined its relationship with the variability of release angle. To obtain accurate finger positions relative to the ball without impeding movement or sensation, an automatic image recognition technology based on deep learning was employed. This approach revealed a noteworthy correlation between the lower middle finger positions prior to acceleration peaks and the reduced variability in release angle, emphasizing the importance of consistent finger positioning during the pre-release phase. This finger positioning of the pitchers with low variability in the release angle is suggested to be robust against the spatial variability of ball movement.

Several studies have investigated factors influencing baseball pitching velocity. However, some measurements require expensive equipment, and some tests need familiarity to perform well. In this study, we adopted field tests executed using affordable equipment in a tryout event for a professional baseball team in Taiwan, 2019. We use half day to test 64 players, and the result of measurement are used to develop a model for predicting pitching velocity of amateur adult pitchers (age: 23.9 ± 2.8 years; height: 180.3 ± 5.9 cm; weight: 81.4 ± 10.9 kg). The measurements and tests in tryout settings should be easy to implement, take short time, do not need high skill levels, and correlate to the pitching velocity. The outcome measures included maximum external shoulder rotation, maximum internal shoulder rotation, countermovement jump (CMJ) height, 20-kg loaded CMJ height, 30-m sprint time, height, age, and weight tests. Multiple regression indicated a moderate correlation between these tests and pitching velocity (adjusted R 2 = 0.230, p = 0.0003). Among the measures, the ratio of loaded CMJ to CMJ, ratio of first 10-m sprint time to 30-m sprint time, and height were significant contributors to pitching velocity. Overall, these measures explained 23% of the variance in the predicted pitching velocity. These field tests can be adopted in tryout events to predict a prospect’s potential and to identify underestimated players. Coaches can obtain an expectation of a pitcher’s performance by comparing his pitching velocity with the predicted value derived from the statistical model presented herein, and the room of growth by comparing his current strength to average strength growth after being drafted and trained with professional coaches.

Although there is a commonly held belief within the baseball community that delivery from the stretch confers more stress at the elbow and shoulder joints than delivery from the windup, there remains little evidence in the literature investigating this hypothesis. This study aimed to help address this gap in the literature by studying both intra-pitcher kinematic sequence variability, and intra-pitcher joint torque variability when throwing from the windup vs. the stretch. We hypothesized that 1) each pitchers’ kinematic sequence would remain similar whether throwing from the windup or stretch, and 2) Kinematic sequence would influence peak arm torque more than delivery method. This cross-sectional 3D biomechanical study included 88 pitches thrown by ten (6 collegiate, 4 high school) pitchers with a mean age of 17.60 ± 2.63 years. Pitch velocity, throwing shoulder/elbow torques and the kinematic sequence of each pitch utilizing segmental peak angular velocities were captured. No statistically significant differences in ball velocity (p = 0.17), peak shoulder external rotation torque (p = 0.80), shoulder extension torque (p = 0.97), or elbow valgus torque (p = 0.83) were found between delivery approaches. Three primary kinematic sequences were identified. Shoulder external rotation torque [F(53,2) = 10.992, ɳ2 = .293, p < 0.00], shoulder extension torque [F(53,2) = 15.517, ɳ2 = .369, p < 0.00] and elbow valgus torque [F(53,2) = 9.994, ɳ2 = .274, p < 0.00] did vary significantly across these three kinematic sequence patterns. Our data suggest that the kinematic sequence influences shoulder and elbow torque more than the delivery approach. Instructing ideal kinematic sequence may be more influential for injury avoidance than delivery method.

Background: The current standard for motion capture data collection in baseball biomechanics is marker-based optical motion capture. Recent advancement in markerless motion capture capabilities has greatly improved accessibility to in-game, high-precision motion capture data, but specific values may differ from markered systems, necessitating separate normative values. For future data comparison, reference data are needed. Purpose: To describe common kinematic variables in baseball pitching using an in-game markerless motion capture system. Study Design: Descriptive laboratory study. Methods: Kinematic data were collected in 2 collegiate stadiums in the National Collegiate Athletic Association Division I Southeastern Conference using 8 synchronized 300-Hz KinaTrax cameras placed around the playing field. The in-game biomechanics data of 51 pitchers from 5 different teams during the 2023 season were collected; only pitchers with >3 outings during the season were included. After averaging all available fastballs thrown in the first inning of a game, we analyzed triaxial trunk (rotation, flexion, and lean) and pelvic rotation angles in the global reference frame, as well as shoulder rotation, shoulder horizontal abduction, shoulder abduction, elbow flexion, stride knee flexion, and hip-shoulder separation. Results: A total of 509 fastballs were analyzed. Mean fastball velocity was 40.98 m/s (91.5 mph), with a vertical break of 42.0 ± 10.6 cm (16.5 ± 4.2 inches) and a horizontal break of 28.0 ± 11.1 cm (11.0 ± 4.4 inches). Mean stride length was 1.41 ± 0.08 m. The mean arm slot was 59.4°± 9.3°, and the mean upper arm slot was 72.9°± 9.3° at ball release. Additional normative time-series data are presented for commonly analyzed variables, and discrete metrics are provided at distinct time points in the pitching cycle. Conclusion: These data fill a need for separate norms for in-game, markerless motion capture performance metrics, and biomechanics for collegiate baseball pitchers to be used as reference values for researchers, coaches, and clinicians. Clinical Relevance: Clinicians should use these results as reference values to contextualize injury mechanisms and the injury-performance trade-off. These data will also educate clinicians on what athletes’ bodies must do to perform when constructing plans of care and return-to-play timelines.

The purpose of this study was to clarify differences in the kinematic and kinetic profiles of the trunk and lower extremities during baseball pitching in collegiate baseball pitchers, in relation to differences in the pitched ball velocity. The subjects were 30 collegiate baseball pitchers aged 18 to 22 yrs, who were assigned to high- (HG, 37.4 ± 0.8 m·s(-1)) and low-pitched-ball-velocity groups (LG, 33.3 ± 0.8 m·s(-1)). Three-dimensional motion analysis with a comprehensive lower-extremity model was used to evaluate kinematic and kinetic parameters during baseball pitching. The ground-reaction forces (GRF) of the pivot and stride legs during pitching were determined using two multicomponent force plates. The joint torques of hip, knee, and ankle were calculated using inverse-dynamics computation of a musculoskeletal human model. To eliminate any effect of variation in body size, kinetic and GRF data were normalized by dividing them by body mass. The maxima and minima of GRF (Fy, Fz, and resultant forces) on the pivot and stride leg were significantly greater in the HG than in the LG (p < 0.05). Furthermore, Fy, Fz, and resultant forces on the stride leg at maximum shoulder external rotation and ball release were significantly greater in the HG than in the LG (p < 0.05). The hip abduction, hip internal rotation and knee extension torques of the pivot leg and the hip adduction torque of the stride leg when it contacted the ground were significantly greater in the HG than in the LG (p < 0.05). These results indicate that, compared with low-ball-velocity pitchers, high-ball-velocity pitchers can generate greater momentum of the lower limbs during baseball pitching. Key pointsHigh-ball-velocity pitchers are characterized by greater momentum of the lower limbs during pitching motion.For high-pitched-ball velocity, stabilizing lower limbs during pitching plays an important role in order to increase the rotation and forward motion of the trunk.Computation of the lower-extremity kinetics and measurement of lower-extremity strength may help clarify the role of muscle strength in determining knee and hip function in baseball pitching.

Background: Increased shoulder distraction force during a baseball pitch may make a pitcher susceptible to rotator cuff or glenohumeral labral injuries. A precursor to a pitching injury may be pain experienced in the throwing arm. Purpose: To (1) compare peak shoulder distraction (PSD) forces in youth baseball pitchers with and without upper extremity pain when throwing a fastball and (2) assess if PSD forces across trials differ between pain and pain-free groups. Study Design: Controlled laboratory study. Methods: A total of 38 male baseball pitchers aged 11 to 18 years were separated into a pain-free group (n = 19; mean age, 13.2 ± 1.7 years; mean height, 163.9 ± 13.5 cm; mean weight, 57.4 ± 13.5 kg) and a pain group (n = 19; mean age, 13.3 ± 1.8 years; mean height, 164.9 ± 12.5 cm; mean weight, 56.7 ± 14.0 kg). Pitchers in the pain group indicated that they experienced pain in their upper extremity while throwing a baseball. Pitching mechanical data from 3 fastballs per pitcher were recorded with an electromagnetic tracking system and motion capture software. The mean PSD (mPSD) was calculated as the mean PSD of 3 pitches per pitcher, the trial with the highest recorded PSD was determined as the maximum-effort PSD (PSDmax), and the PSD range (rPSD) was defined as the difference of the PSD force of the trial with the highest PSD and the lowest PSD for each pitcher. The PSD force was normalized to the pitcher’s body weight (%BW). Pitch velocity was also recorded. Results: The mPSD force was 114%BW ± 36%BW for the pain group and 89%BW ± 21%BW for the pain-free group. Pitchers in the pain group exhibited a significantly higher PSDmax force (t 30.548 = 2.894; P = .007) and mPSD force (t 29.231 = 2.709; P = .009) compared with those in the pain-free group. There were no significant between-group differences in the rPSD force or pitch velocity. Conclusion: The normalized PSDmax force was higher in pitchers who experienced pain while throwing fastballs compared with pitchers who were pain-free while throwing. Clinical Relevance: Baseball pitchers who experience pain in their throwing arm are likely to have higher shoulder distraction forces. Improvement in pitching biomechanics and corrective exercises may assist in the mitigation of pain while pitching.

In the game of softball, the batter should possess the necessary skills to hit the ball toward various directions with high initial speed. However, because various factors influence each other, there are limitations to the range that can be controlled by the batter’s skill. This study was aimed at extracting the impact characteristics associated with the launch speed/direction and batted ball spin and clarifying the important skills required to improve the batter’s hitting performance. In our experiments, 20 female softball players, who are members of the Japan women’s national softball team, hit balls launched from a pitching machine. The movements of the ball and bat before, during, or after the impact were recorded using a motion capture system. Stepwise multiple regression analysis was performed to extract factors relating the side spin rate. The undercut angle (elevation angle between the bat’s trajectory and the common normal between the ball and bat: ΔR 2 = 0.560) and the horizontal bat angle (azimuth of bat’s long axis at ball impact: ΔR 2 = 0.299) were strongly associated with the side spin rate (total R 2 = 0.893, p < 0.001). The undercut angle in opposite-field hitting was significantly larger than that in pull-side hitting ( p < 0.001). The side spin rate was associated with the undercut angle because the bat’s distal (barrel) side inclined downward (–29.6 ± 8.7°) at impact. The ball exit velocity was higher when it was hit at a smaller undercut angle (R 2 = 0.523, p < 0.001). Therefore, it is deemed desirable to focus on maximizing the ball exit velocity rather than ball spin because the ball–bat impact characteristics vary inevitably depending on the launch direction. Meanwhile, the use of the ball delivery machine and the slower pitched ball are the limiting factors in the generalization of the findings.

This paper proposes a baseball pitching action recognition algorithm based on a spatiotemporal graph convolutional neural network and constructs an error action correction algorithm on this basis. The dynamic skeleton model ST-GCN is used to combine the positional information of human movement with the temporal dynamic information. The action contour sequence is extracted to determine the funding for the erroneous action. Finally, the machine learning method is used to realize the adaptive corrective analysis of the erroneous action. Example analysis shows that the action correction algorithm proposed in this paper improves the recognition accuracy by 20.78%, 16.67%, and 9.11%, 9.73% in the two datasets, and the pitching accuracy of the experimental group is 12.5% higher than that of the control group, and the standardized degree score of the pitching technical action is 1.1 points higher than that of the control group. Therefore, the practical effectiveness of the pitching action identification and correction method in this paper has been effectively verified.

The aim of this systematic review was to critically review and synthesise the findings from primary studies on pitching mechanics and performance of healthy adult baseball pitchers. Systematic review with meta-analysis. Eight English- and Japanese-language databases were systematically searched from inception to 22nd July 2022. In total, 29 descriptive biomechanical studies were included. Overall study quality was moderate. In subgroup analyses, professional pitchers showed significantly higher stride length, peak shoulder internal rotation velocity, peak shoulder proximal force and ball velocity compared to collegiate pitchers. Conversely, collegiate pitchers were found to have significantly higher peak pelvis rotation velocity. Available normative data suggested potentially heterogeneous pitching mechanics and performance between professional and collegiate pitchers. However, the findings in this review should be interpreted cautiously. Since statistical heterogeneity was significant within most data sets, more detailed subgroup analyses are required. Additionally, more high-quality studies utilising measurement systems with established reliability are required to obtain accurate data in baseball pitching mechanics and performance.