A precise and reliable quantification of training load (TL) for each individual is given great importance in sports training. It could contribute to a more accurate assessment of how an athlete is responding to a prescribed training and assist in subsequent changes to the training program to avoid overtraining and improve training efficiency. The evaluation of TL varies between sports and training modalities, especially in basketball, football, and other team sports, which demand a hybrid energy supply, and which involve frequent physical contact, complex technical actions, varied tactics, and diverse training methods, making precise TL difficult to assess.

Currently, there are multiple methods of TL quantification in basketball. The external load measures include distance, speed, time-motion analysis, accelerometer load, and global positioning system (GPS) parameters, while measures such as heart rate (HR), blood lactate, and ratings of perceived exertion (RPE) are commonly used to assess internal load (Piedra et al., 2021). However, different TL evaluation tools and indicators have their advantages and limitations. For example, GPS devices provide a real-time measure of an athletes position, speed, and distance which is widely used in outdoor team sports, but it is susceptible to a variety of interferences owing to the low signal power, and thus presents poor measurement reliability during short high-speed movements and multidirectional exercises (Crang et al., 2021). Video techniques not only provide duration, distance, and speed during exercise but also accurate analysis of complex sport-specific technical actions (Carling et al., 2008). However, it is incapable of real-time measurements and is time-consuming to process. Traditional biochemical markers such as blood lactate concentration can accurately reflect the body’s response to exercise stimuli, but measurement precision is subject to errors derived from the blood sampling site, time, and technique. HR, as a non-invasive and convenient method, is commonly used for monitoring exercise intensity in the field, but studies have shown that HR and intensity of aerobic exercise are linearly correlated, while in interval and resistance training, the relationship between these two indicators is not linear (Borresen and Lambert, 2009). Thus, monitoring basketball training and competition load with HR alone may underestimate exercise intensity (Narazaki et al., 2009). RPE has been developed and validated as a simple and effective method to assess the internal TL of athletes in different sports. However, athletes may exaggerate their RPE in some situations, considering that subjective sensations are influenced by many factors, including physical health, energy supply, and psychological state (Fox et al., 2017). In addition, the RPE value may be different for different training purposes. For example, Lupo et al. found that the intensity of taekwondo competition determined by RPE which tends to be lower than that measured by HR-based methods. Conversely, RPE was more consistent with HR-measured TL in training sessions (Lupo et al., 2017). The training impulse (TRIMP) is a measure that reflects the comprehensive response of the body to internal and external load stimuli through training volume (duration) and intensity (average HR). However, some studies have shown that TRIMP cannot distinguish between continuous and interval training of the same duration and average HR, which limits its application in short-duration, high-intensity interval training (Buchheit and Laursen, 2013). Therefore, in practice, there is no real gold standard for TL quantification.

Recently, heart rate variability (HRV) has been widely used to monitor training as a non-invasive quantitative assessment of cardiac autonomic tone and balance, which is sensitive to changes of psychophysiology and fatigue of athletes under external stimulation (Deus et al., 2019; Schneider et al., 2019; Lukonaitiene et al., 2020). The root mean square of successive differences between beat intervals (RMSSD), as a time-domain marker of HRV, is very sensitive to the short-term and rapid regulation of vagus nerve activity in the autonomic nervous system (Rave et al., 2020). During exercise, the release of norepinephrine triggers sympathetic activation, which leads to increased cardiac contractility and conduction of electrical signals to meet the physiological needs of exercise. At the same time, the vagus nerve is inhibited and RMSSD declines dramatically. When at rest, the parasympathetic vagus nerve promotes recovery by releasing acetylcholine, slowing down the HR, reducing cardiac contractility, and inhibiting cardiac electrical signaling speed, and thus sympathetic nerve activity is inhibited, and RMSSD increases significantly. RMSSD declines sharply during exercise and then becomes reactivated after exercise, which can accurately reflect the stress response of different individuals to TL (Shaffer et al., 2014; Gordan et al., 2015; Grassler et al., 2021). Saboul et al. proposed the HRV index (TL_HRV) as a new TL quantification tool based on RMSSD related research and verified the effectiveness of TL_HRV in TL assessment of continuous and interval running (Saboul et al., 2016). Zhao et al. further demonstrated that TL_HRV was in line with TRIMP in evaluating the TL of continuous exercise (Zhao et al., 2018). However, there is still uncertainty whether TL_HRV could effectively quantify TL in high-intensity interval events such as basketball and football.

Thus, the purpose of this investigation was to verify whether TL_HRV in assessing the TL of basketball is in line with TRIMP, RPE, speed, and distance during five ball-drills (half-court 1V1, 2V2, 3V3, 4V4, and 5V5) or not. It was hypothesized that TL_HRV was able to differentiate loads between the five basketball drills, and it showed a good agreement and correlation with the validated internal and external load monitoring indicators.

The purpose of this investigation was to verify the effectiveness of TL_HRV in quantifying the TL of basketball, by exploring the relationship between TL_HRV and internal and external load measures during five ball-drills in trained athletes. It should be noted that this is the first study to evaluate the TL of basketball by TL_HRV. The current study found that TL_HRV was able to differentiate loads between various basketball drills, and it showed good agreement and correlation with TRIMP, RPE, and running speed.

The results showed that TL_HRV of 1V1, 2V2, 3V3, 4V4, and 5V5 decreased gradually with the increase of the number of athletes in each group. This may be attributed to the fact that the number of players affected the physical demands of basketball, which stimulated the parasympathetic nerves to innervate the arteries and myocardium by releasing acetylcholine to produce corresponding adaptive physiological changes. This result is in line with those of previous observations. For example, Castagna et al. found that the fewer players on the court, the higher the average heart rate and blood lactate concentration of athletes during 5V5, 3V3, and 2V2 drills (Castagna et al., 2011). The percentage of maximal HR and peak VO₂ has been reported to increase by 3%–8% and 10%, respectively, with a decrease in the number of players in basketball small sided games (Conte et al., 2016). In fact, the number of athletes on the court causes differences in factors such as the frequency of techniques, the rhythm of the game, and the relative court area per player (Fernandez-Espinola et al., 2020). Specifically, the fewer players on the court, the more opportunities each player has to touch the ball, and it can increase the individual physical contacts, technical actions, and the relative court area per player, thereby changing the physical requirements of the exercise, leading to the differences in TL of the five drills. Conversely, Wang et al. showed that the TL of 4V4 and 3V3 were higher than 5V5, 2V2, and 1V1 in a half-court game (Wang et al., 2018). This may be related to the different amount of work each group was carrying out in their experiment. The training durations of 1V1, 2V2, 3V3, 4V4, and 5V5 were 9, 11, 9, 13, and 16 min, respectively. Comparing the intensity of different forms of exercise under different training durations may cause some deviation in research results.

The results of the analysis clearly demonstrated the correlation between the two internal TL variables and TL_HRV; similar results were found in previous research related to various indicators of HRV. Zhao et al. have investigated correlations between TL_HRV and TRIMP in female soccer athletes; they investigated TL_HRV at 60% and 85% maximal speed of the graded maximal test and found moderate correlations between TL_HRV and TRIMP (Zhao et al., 2018). Moderate negative correlations between post5 RMSSD and TRIMP were reported by González-Fimbres in the high-intensity interval and low-intensity continuous running of youth triathletes (González-Fimbres et al., 2020). Orellana et al. observed moderate correlations between post10 RMSSD and TRIMP by examining 14 healthy men who ran at ventilatory thresholds (VT1 and VT2) and maximal aerobic speed (Orellana et al., 2019). As for the relationship of HRV and RPE, moderate correlations between morning HRV and session ratings of perceived exertion (sRPE) of the previous day were found by Sartor et al. in young male elite gymnasts (Sartor et al., 2013). Kaikkonen et al. illustrated larger negative correlations between RPE and HF, and TP in interval running exercises with different intensities and durations (Kaikkonen et al., 2012). The studies above suggest that HRV is related to TRIMP and RPE in various sports and exercise, a possible explanation for this might be that the three indicators are based on the relationship between HR and exercise intensity, which can reflect the change of cardiac function with TL during exercise.

In the present study, a significant relationship was found between TL_HRV and speed. Interestingly, the distance was not related to TL_HRV. Similar conclusions were reported by Seiler et al., who explored the effect of exercise duration on RMSSD in endurance athletes, finding that post-exercise RMSSD was not affected by duration, despite a two-fold increase in duration (Seiler et al., 2007). Research by Kaikkonen et al. showed that HRV recovery was slower with increased exercise intensity, while doubling distance had no effect on HRV recovery (Kaikkonen et al., 2007). Saboul et al. also did not find a correlation between TL_HRV and exercise volume in long-distance male runners and argued that regardless of the type of exercise, exercise intensity was the main factor affecting baseline HRV and recovery of the body post-exercise (Saboul et al., 2016). Based on the above evidence, we believe that high-intensity training may delay HRV recovery, and the higher the exercise intensity, the lower the RMSSD in the same recovery time. Moreover, it is also suggested that high-intensity exercise (above the second ventilatory threshold) seems to have a greater effect on parasympathetic regulation of cardiac activity compared with exercise at very low intensity (below the first ventilatory threshold) (Kaikkonen et al., 2010). This is partly explained by the fact that low-intensity exercise mainly relies on the aerobic energy system, when blood lactate concentrations generally do not exceed the resting value. In contrast, high-intensity exercise is primarily powered by the anaerobic glycolytic metabolic system, which leads to a large increase in blood lactate concentrations during exercise, triggering metabolic reflexes that inhibit parasympathetic activity, and in turn lower HRV. Collectively, these physiological mechanisms create intuitive rationales for the significant associations between running speed and TL_HRV found in our analyses (González-Fimbres et al., 2020).

The aim of sports training is to disturb the body’s homeostasis by applying appropriate stimuli to the body, including sympathetic and vagal activity. Monitoring whether such stimulation is suitable for the body using precise and convenient methods or tools is a milestone in athletic training. Although there is no “gold standard” to quantify TL, TRIMP may serve as a reference parameter that partially reflects the body’s physiological response to stress. In addition, some studies found that RPE was correlated with the cardiorespiratory and metabolic demands of the body during exercise, and therefore had been also used as a reference index for TL evaluation. However, the measurement of TRIMP cannot distinguish between continuous and interval training of the same duration and average HR, which limits its application in short-duration, high-intensity interval training (Buchheit and Laursen, 2013). RPE is simple and effective in training practice, but it is subjective and cannot directly reflect the athletes’ physiological response during training (Kaikkonen et al., 2012). In pursuit of an objective and convenient physiological indicator to quantify TL in basketball, this study investigated the emerging TL_HRV parameter. The results of the study were basically in line with the expected hypothesis, that is, the ball-drills of different intensity did show some differences under the quantification of TL_HRV. Accordingly, we cautiously conclude that TL_HRV could be used as markers of training load and psychophysiological status of male basketball players, as it can provide unique and deep insights into athletes’physiology response compared to TRIMP, RPE, speed, and distance. Furthermore, it is easy for practitioners to obtain the RMSSD indicators and the relevant data analysis for free from some HR monitoring systems. From a practical point of view, we encourage practitioners to monitor changes in TL_HRV during a training session, and may even attempt to monitor it continuously over a training period, which could contribute to a more objective and accurate assessment of how an athlete is responding to a prescribed training and assist in subsequent changes to the training program to avoid overtraining and improve training efficiency.

In this study, the correlation analysis between TL_HRV and internal load indicators (TRIMP and RPE) and external load indicators (running distance and speed) was used to preliminarily verify that TL_HRV can be used to evaluate TL in basketball. However, the study was limited by the absence of analysis concerning biochemical indicators such as blood lactate concentration. Therefore, in future research, it would be beneficial to correlate TL_HRV with more physiological and biochemical indicators to further confirm its accuracy in basketball TL evaluation.

Furthermore, the TL_HRV formula is based on pre5 RMSSD, post5 RMSSD, and post30 RMSSD. We should acknowledge that it takes a certain amount of time to collect RMSSD data at three time points, which is more complicated than collecting RPE data for daily use. In addition, this formula has not been extensively tested across different sports and training modalities, and its applicability to load monitoring in other forms of exercise is still unclear. From this perspective, future research could focus on demonstrating the validity of TL_HRV in different sports and training modalities.

Based on the present study, we cautiously conclude that TL_HRV may serve as an objective and rational measure for monitoring TL in basketball players. In basketball training practice, it is suggested that coaches can judge the physiological adaptation of athletes to training load according to changes in TL_HRV after training sessions, which will help practitioners to adjust training plans and meet specific training goals.