Micro electro-mechanical systems (MEMS) are used to record training and match play of intermittent team sport athletes. Paired with estimates of internal responses or adaptations to exercise, practitioners gain insight into players’ dose–response relationship which facilitates the prescription of the training stimuli to optimize performance, prevent injuries, and to guide rehabilitation processes. A systematic review on the relationship between external, wearable-based, and internal parameters in team sport athletes, compliant with the PRISMA guidelines, was conducted. The literature research was performed from earliest record to 1 September 2020 using the databases PubMed, Web of Science, CINAHL, and SportDISCUS. A total of 66 full-text articles were reviewed encompassing 1541 athletes. About 109 different relationships between variables have been reviewed. The most investigated relationship across sports was found between (session) rating of perceived exertion ((session-)RPE) and PlayerLoad™ (PL) with, predominantly, moderate to strong associations (r = 0.49–0.84). Relationships between internal parameters and highly dynamic, anaerobic movements were heterogenous. Relationships between average heart rate (HR), Edward’s and Banister’s training impulse (TRIMP) seem to be reflected in parameters of overall activity such as PL and TD for running-intensive team sports. PL may further be suitable to estimate the overall subjective perception. To identify high fine-structured loading—relative to a certain type of sport—more specific measures and devices are needed. Individualization of parameters could be helpful to enhance practicality.

Marker-based (MB) motion capture systems face challenges like marker loss, soft tissue artifacts, and manual processing. Markerless (ML) motion capture systems might be particularly advantageous for capturing dynamic, non-linear movements like change of direction (COD) movements. The agreement between MB and ML systems for lower extremity joint kinematics during COD tasks was evaluated. Nineteen athletes performed cutting movements in 5 directions at 3 intensities. Data was captured using infrared and video cameras. Joint angles were computed, and the agreement was assessed using prediction band and extended Bland-Altman (BA) methods. Knee joint angles showed the smallest random and systematic errors (bias = 4.34°, area = 3102.57 deg·stance%; BA: bias = 3.34°, limits of agreement (LoA) = ± 11.38°) compared to ankle and hip joint angles. Flexion/extension angles displayed a smaller random error (area = 2919.74 deg·stance%; LoA = ± 10.71°) compared to ab-/adduction (area = 3477.70 deg·stance%; LoA ± 12.16°) and internal/external rotation angles (area = 4630.68 deg·stance%; LoA = ± 15.75°). Slower and straight-line movements demonstrated stronger agreement than sharper, non-linear directions. These findings provide valuable insight for researchers and practitioners when placing ML data in the context of existing data, with particular care when considering highly dynamic, non-linear movements or internal/external rotation angles.

Markerless (ML) motion capture is a practical alternative to marker-based (MB) systems, yet its ability to estimate joint kinetics during team sport-specific movements remains unclear. This study evaluated the agreement between ML and MB lower extremity joint moment estimates during change-of-direction (COD) tasks. Nineteen team sport athletes performed straight running and four COD movements (45-, 90-, 135-, 180-degrees) at different speeds. Joint moments at the hip, knee, and ankle were computed from synchronized force-plate data and either ML video data (Theia3D) or MB motion-capture data. Agreement was assessed using an extended Bland–Altman approach with linear mixed-effects models for peak joint moments and bootstrapped functional prediction bands for the entire stance phase. Substantial random and systematic differences were observed between systems exceeding the defined threshold of ≤ 10% in relative limits of agreement (LoAs) and in average MB signal range × 100% stance. LoAs for peak joint moments ranged from ± 0.38 Nm/kg at the ankle to ± 1.21 Nm/kg at the hip, indicating that difference between systems was of similar magnitude to, or larger than, the underlying MB-derived peak moments. Relative random differences ranged from 92 to 128% across movement planes. Agreement did not show a consistent pattern across movement directions or direction-specific speed categories. These findings indicate that ML- and MB-derived joint kinetics cannot be used interchangeably.