ABSTRACT
This study aimed to describe the physical demands of American football players using novel performance analysis techniques. Heart rate (HR) and accelerometer-based activity levels were observed across two pre-season scrimmages in 23 Division I collegiate football players (age: 19 ± 1 y, height: 1.90 ± 0.06 m, weight: 116.2 ± 19.4 kg). Data were analyzed using a MATLAB program and inter-rater reproducibility assessed using inter-class correlations (ICC). Players were analyzed by side (offense/defense) and position (skill/non-skill). Performance variables assessed in bursts of activity included burst duration, HRmean and HRmax (bpm), and mean activity (vector magnitude units [vmu]). Exercise intensity was categorized as time spent in % HRmax in 5% increments. The burst duration (8.1±3.9 min, ICC = 0.72), HRmean (157 ± 12 bpm, ICC = 0.96) and mean activity (0.30 ± 0.05 vmu, ICC = 0.86) were reproducible. HRmean (p = 0.05) and HRmax (p = 0.001) were greater on defense. Offense spent more time at 65-70% HRmax (p = 0.01), 70-75% HRmax (p = 0.02) while defense spent more time 90-95% HRmax and ≥95% HRmax (p = 0.03). HRmean (p = 0.70) and HRpeak (p = 0.80) were not different between positions across both sides. Skilled players demonstrated greater mean activity (p = 0.02). The sport-specific analysis described HR and activity level in a reproducible manner. Automated methods of assessing HR may be useful in training and game time performance but ultimately provides support to coaching decision making.
