Comparison between RPE- and accelerometer-based training loads during semi-professional basketball training

Introduction: To date, internal measures have largely been used to quantify training load (TL) in basketball players. Specifically, internal TLs calculated using player ratings of perceived exertion (RPE) have been reported to strongly relate with those using heart rate (HR) responses. However, the relationship between RPE- and HR-based TLs has been shown to vary with different training types in other team sports. Further, various limitations have been reported when using HR to evaluate high-intensity intermittent work. As such, quantification of RPE-based TLs relative to external measures across different training types warrants investigation in basketball players. Thus, the purpose of this study was to explore the relationship between TL calculations derived from RPE and accelerometer measures across different types of basketball training. Methods: Eight state-level Australian male basketball players (mean +/- SEM, age: 27.3 +/- 2.6 yr; stature: 189.4 +/- 2.3 cm; body mass: 94.8 +/- 3.5 kg) were monitored during basketball-specific off-court interval training and on-court conditioning sessions across the preparatory phase of the annual plan. Approximately 30 min following each session, players reported their perceived intensity (Borg CR-10 scale) of the overall session to calculate session-RPE TL (RPE x duration (min)). Additionally, triaxial accelerometers were secured to players on the posterior surface of the torso at the sternal level via an elastic strap, and sampled at 100 Hz across the entire training session. Accelerometer-based TL was determined using previously formulated vector magnitude calculations accumulated over the length of the session. Mean session TLs were calculated using both RPE- and accelerometer-based calculations for each training type. Pearson correlation analysis was conducted to determine the relationship between methodologies. Results: Moderate to strong correlations were observed between RPE- and accelerometer-based TLs (in arbitrary units) during off-court (n=25; RPE: 262 +/- 9; accelerometer: 4772 +/- 39; r=0.71, p<0.001) and on-court training sessions (n=14; RPE: 365 +/- 7; accelerometer: 5157 +/- 30; r=0.57, p<0.05). Discussion: The present data supports the use of RPE-based TLs in comparison with accelerometer-based methods in basketball players. Stronger correlations were observed between methods during off-court sessions, possibly due to a dissociative increase in RPE during on-court drills from greater static exertion, upper-body activity, and directional changes. The monitoring of TLs using player RPE alone or in conjunction with accelerometers might allow coaching and training staff to implement more effective training plans to optimise physiological adaptation and avoid overtraining in basketball players.