(18) AGREEMENT BETWEEN MOTION TRACKING AND VIDEO ANNOTATION SOFTWARE FOR MONITORING BARBELL VELOCITY AND POWER DURING HIGH INTENSITY FUNCTIONAL TRAINING

Thursday, July 13, 2023

2:00 PM – 3:30 PM East Coast USA Time

Location: Milano III-IV

Poster Presenter(s)

Ashley Hines

Graduate Research Assistant
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Kennesaw, Georgia, United States

Author(s)

JF

Jacob Fanno

Student
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Kennesaw, Georgia, United States
Christopher Staples

Graduate Assistant
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Acworth, Georgia, United States
Jacob Grazer, PhD, CSCS

Program Director: MS in Exercise Science
Kennesaw State University
Atlanta, Georgia, United States
TE

Tiffany Esmat

Professor/Chair
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Kennesaw, Georgia, United States
JM

John McLester

Professor
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Kennesaw, Georgia, United States
Gerald T. Mangine, PhD

Associate Professor
Exercise Science and Sport Management, Kennesaw State University, Kennesaw, Georgia
Kennesaw, Georgia, United States

Abstract Details: High-intensity functional training (HIFT) involves vigorous multimodal circuit training with auto-regulated rest periods. Performance is typically gauged by how quickly work is completed, and to this end, dependent on finding the fastest sustainable pace for each exercise prescribed within a workout. However, objectively finding and precisely tracking ideal pacing is nearly impossible without technological assistance. One might record workouts and then utilize video annotation software (VAS) to quantify kinetics, but this method is tedious and limited by camera quality and standardization procedures. In contrast, recent advances in 3D motion tracking systems (MTS) can provide real-time feedback with fewer procedural restrictions, though technologies may be limited to specific exercises and at times, can fail to detect repetitions. Data loss might be avoided by having VAS serve as back-up for when repetitions are not detected. However, agreement between technologies must first be established.

Purpose: To determine agreement between VAS and MTS for monitoring barbell velocity and power during a 5-minute bout of HIFT.

Methods: A sub-sample of seven, male HIFT-trained athletes (28.8 ± 7.2 years, 172.9 ± 9.2 cm, 83.2 ± 17.0 kg) from a larger investigation examining the effect of a pre-workout supplement on HIFT workout performance was used for this study. Over a 4-week period, participants randomly completed a 5-minute repeated circuit of 9-calorie rowing, six barbell thrusters at 43.1 kg, and three 0.61 m box jumps after consuming supplement or placebo. All workouts were recorded by tablet camera with average concentric barbell velocity (V) and power (P) during thrusters being specifically tracked by MTS (PERCH, Catalyft Labs, Inc., Cambridge, MA, USA). Post-exercise, VAS (Kinovea v.0.9.5, Free Software Foundation, Inc., Boston, MA, USA) was used to also quantify average concentric V and P for all completed thruster repetitions during both 5-minute workout sessions. Subsequently, quantified kinetics for all complete sets (i.e., six detected repetitions from both MTS and VAS) were organized (n = 321 sets), then a random sample of 219 sets (~68%) was selected for agreement analysis.

Results: Differences and relationships between raw scores obtained by MTS and VAS were initially assessed by paired samples t-tests and Pearson's correlation analysis. Though related (r = 0.36 - 0.40, p < 0.001), V and P were significantly (p < 0.001) different between MTS (V: 1.21 ± 0.16 m/s; P: 510 ± 69 W) and VAS (V: 1.48 ± 0.33 m/s; P: 626 ± 125 W). Negative Pearson’s correlations between the averages and differences (between MTS and VAS) in for V (r = -0.64, p < 0.001) and P (r = -0.63, p < 0.001), as well as associated Bland-Altman plots, indicated lower MTS estimates of V (mean bias = -0.27 m/s, 95% C.I. = -0.89 to 0.34 m/s, CV = -114%) and P (mean bias = -117 W, 95% C.I. = - 361 to 128 W, CV = -107%) compared to VAS, and that these were not consistent across the range of velocities (0.81 - 1.80 m/s) and power (342 - 759 W) observed across all repetitions.

Conclusion: Compared to MTS, VAS overestimates V and P. Differences are exacerbated with faster or more powerful repetitions. PRACTICAL APPLICATION: Simultaneous utilization of MTS and VAS, with the latter serving as back-up for undetected repetitions, would not be a suitable solution for tracking barbell thruster kinetics during a HIFT workout.

Acknowledgements: This study was funded by Shifted LLC, Eugene, OR, USA.