Heart Rate and Activity Measured during Volleyball Competition

using Wearable Technology

Jasper Gielen

, Elise Mehuys

, Kris Eyckmans

and Jean-Marie Aerts

M3-BIORES, Division Animal and Human Health Engineering, Department of Biosystems, KU Leuven,

Kasteelpark Arenberg 30, 3001 Heverlee, Belgium

TopVolleyBelgium, Beneluxlaan 22, 1800 Vilvoorde, Belgium

Keywords: Volleyball, Heart Rate, Activity, Competition, Wearables, Physiological Monitoring.

Abstract: Volleyball is characterised by intervals of high-intensity gameplay mixed with periods of relative rest.

Monitoring the athletes’ physiology during competition allows us to study the changes in exercise intensity

throughout a game. In this study, eight elite male volleyball athletes measured their heart rate and activity

during multiple games of the regular season in the Belgian Liga A and B using wearable technology. The data

show a significant decrease in the heart rate for set 1 to 4, from 79.1 %HR

max

to 73.9 %HR

max

. For activity, a

decreasing trend is visually observed, but the difference is only significant for set 1 compared to the other

sets. Finally, the performance did not vary significantly over the course of the different sets.

1 INTRODUCTION

Volleyball is a team sport with a highly variable and

dynamic nature (Künstlinger et al., 1987). The teams

switch back and forth between offence and defence

within a matter of seconds. Moreover, scoring can be

the result of a single attack or a long-lasting, high-

intensity rally. In addition to the dynamics of the

game, players take up different roles and accordingly

require a different skillset (Sheppard et al., 2013). To

illustrate, the libero is a defensive specialist and

cannot take part in offence, the setter leads the offence

and decides which hitter will attack, and the outside,

middle and opposite hitters each have their own

position for attacking. Studying volleyball athletes’

physiology is interesting to identify the efforts that

are required during a game and analyse how this

changes over the course of the different sets.

From a physiology point of view, volleyball has

traditionally been described as a high-power,

predominantly anaerobic sport (Van Heest, 2003).

Due to the rules and structure of the game, athletes

experience intervals of intense exercise, but also have

the opportunity to recover in between these intervals.

During these intense exercise intervals, the athletes’

body generates energy primarily from creatine

phosphate stored in the muscle cells and from

anaerobic glycolysis. During recovery periods,

intracellular stores of creatine phosphate are

replenished through aerobic pathways (Van Heest,

2003). It has been estimated that the overall energy

demands of the sport are met by a combination of all

three energy-producing pathways in the following

proportions: the creatine phosphate system (40%),

anaerobic glycolytic system (10%) and aerobic

metabolism (50%) (Gionet, 1980).

Heart rate monitoring is a popular tool for

quantifying the internal load in athletes (Buchheit,

2014). This approach is based on the linear

relationship between steady-state work rate and heart

rate (Hopkins, 1991; Arts & Kuipers 1994). Despite

the abundance of tools for heart rate monitoring, data

during volleyball competition is limited.

Furthermore, the data that are available indicate a

high variability. In the research of González et al.

(2005), a mean value for the heart rate of 148 bpm

was observed for the principal central players during

their time on the court. This decreased to 124 bpm for

the periods off the court. For the liberos, mean values

of 137 and 131 bpm were noted for the time on and

off the court respectively. In beach volleyball, a mean

heart rate of 146 bpm was noted for a 3-set match

(Jimenez-Olmedo et al., 2017).

In this study, we aimed at analysing the evolution

of the heart rate over the different sets. A similar

analysis is performed with accelerometer data to

identify changes in activity throughout the match.

212

Gielen, J., Mehuys, E., Eyckmans, K. and Aerts, J.

Heart Rate and Activity Measured during Volleyball Competition using Wearable Technology.

DOI: 10.5220/0010047802120216

In Proceedings of the 8th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2020), pages 212-216

ISBN: 978-989-758-481-7

This way, we aim to gain insights into the exercise

intensity during volleyball competition. Finally, the

performance of the athletes was analysed.

2 MATERIAL AND METHODS

2.1 Experiment Design and

Participants

Eight male elite volleyball athletes participated in this

study. Four athletes collected data during games in

the regular season of 2017-2018. They played

together on a team in the Belgian Liga A (1

division). The other four athletes were monitored

during the season of 2018-2019 and played on a team

in the Belgian Liga B (2

division). Table 1 shows

the athlete’s age, height, playing position, division

and number of games during which data collected. In

total, 63 unique measurements are included in this

study.

Table 1: Info about the participating athletes.

Pla

e Hei

ht Position Division #

ames

1 19 193 c

Outside Li

a A 8

2 21 197 c

Outside Liga A 8

3 19 174 c

Libero Liga A 8

4 32 195 c

Opposite Liga A 4

5 18 191 c

Sette

a B 10

6 16 188 c

Outside Li

a B 9

7 17 194 c

Outside Liga B 9

8 16 200 c

Cente

Liga B 7

Athletes were equipped with a wearable device to

monitor physiological data throughout the entire

game. The measurements were started before the

warm-up and were stopped when the game was

ended. The experimental set-up and data collection

were approved by the social and societal ethics

committee (SMEC) of the KU Leuven (cases ‘G-2017

11 999’ and ‘G-2018 11 1432’).

2.2 Data Collection

The BioHarness 3.0 chest strap (Zephyr Technology,

USA) was used to monitor heart rate and activity

continuously without affecting the athletes’

performance. The device recorded the electrical skin

potential differences at 250 Hz to construct the ECG

signal. A value for the heart rate was automatically

calculated and sampled to 1 Hz. Additionally, this

heart rate signal was checked to match the original

ECG rhythm.

Next to the heart rate, the device captured

acceleration in a three-dimensional way at a

frequency of 100 Hz. The axial accelerometer output

was automatically band-pass filtered to remove non-

human artefacts and gravity. This data was further

processed by taking the mean acceleration for each

axis per second and calculating the root-mean-square

of the three axes. As a result, one signal for the

activity of the athlete is obtained with a sampling

frequency of 1 Hz.

Data regarding the actions and events that

occurred during the competition were captured

through the DataVolley scouting software (Data

Project, Italy). The team’s scouter labelled each

contact with the ball as a serve (S), reception (R),

attack (A), block (B), dig (D), set (E) or free ball (F).

Additionally, raw video files were also available. This

way, the timing of the different sets, the rotations and

the substitutions were noted. In our analysis, we will

take a closer look at the total number of actions that

were performed during the game and the duration

during which a player was on the playing field.

2.3 Data Analysis

2.3.1 Data Processing

First, the data was processed so that only the periods

during which the athlete was part of the game were

retained. This means filtering out substitutions. Since

the libero has a special role and can be replaced

continuously, only long term substitutions were

considered. As González et al. (2005) also indicate,

the time spent off-court for the libero is very short (29

s). Therefore, the libero has to be ready to enter the

field at any moment.

Second, volleyball games are regularly

interrupted by time-outs. During set 1 to 4, technical

time-outs of 60 seconds are automatically imposed

when the leading team reaches a score of 8 or 16.

Furthermore, one or two additional time-outs can be

taken by each team depending on the specific rules of

the competition. Data during the time-outs were also

omitted.

Third, the heart rate data was converted to a

percentage of the maximal heart (%HR

max

) for each

athlete. This was done to account for inter-individual

differences in the range of the heart frequency.

Fourth, the data were processed per set. For each

set, the mean values for the heart rate percentage,

activity, numbers of volleyball actions and minutes

played were calculated.

Heart Rate and Activity Measured during Volleyball Competition using Wearable Technology

213

2.3.2 Statistical Analysis

For the analyses, the data of all athletes and matches

are considered together. The analyses are performed

on the level of the entire test population.

The aim of the statistical analysis was to

determine whether the variables vary significantly

throughout the game. First, a Kruskal-Wallis test was

performed on the data for the different sets. This non-

parametric test, checks the hypothesis that the

distribution for at least one set differs significantly

from another set. Secondly, significant Kruskal-

Wallis tests were followed up with a pairwise

comparison for each combination of two sets using a

pairwise Student’s t-test. For p-values smaller than a

significance level (α) of 0.05, the tests are considered

to be significant. Prior to the Student’s t-test, the

normality of the data was checked.

3 RESULTS

3.1 Heart Rate Throughout the Game

The evolution of the mean value for the %HR

max

per

set is visualised in Figure 1. The bars indicate the

mean value for the distribution of each set and the

whiskers indicate the standard deviation. Visually, it

is clear that the heart rate tends to decrease from set 1

to 4. For set 5, this decreasing trend is not continued.

It has to be noted that both sets 4 and 5 are not

played each game since the match ends when one

team wins three sets. The number of measurements

for set 1 to 5 is respectively 61, 61, 55, 30 and 5.

Figure 1: Evolution of the mean for %HR

max

over the

different sets. The whiskers indicate the standard deviation.

The Kruskal-Wallis test indicates a significant

difference (p = 2.94e-6). Furthermore, Table 2 shows

the p-values that correspond to the pairwise Student’s

t-test for each combination of two sets. The

decreasing trend from set 1 to set 4 is significant. The

mean heart rate for set 5 does not differ significantly

from any other set.

Table 2: P-values for the pairwise Student’s t-test for the

mean heart rate. Values smaller than the significance level

(α = 0.05) are presented in italic.

-value S1 S2 S3 S4 S5

S2 < 0.001

S3 < 0.001 0.00

S4 < 0.001 0.00

0.045

S5 0.198 0.661 0.703 0.905

3.2 Activity Throughout the Game

The evolution of the mean value for the activity per

set is visualised in Figure 2. The bars indicate the

mean value for the distribution of each set and the

whiskers indicate the standard deviation. Visually,

the activity tends to decrease from set 1 to 5.

Figure 2: Evolution of the mean activity over the different

sets. The whiskers indicate the standard deviation.

The result of the Kruskal-Wallis test is significant

(p = 0.003). Additionally, Table 3 shows the p-values

for the pairwise Student’s t-test for each combination.

The tests indicate that the difference in activity is only

significant for set 1 compared to set 2, 3 and 4. For

the other combinations, this is not the case.

Table 3: P-values for the pairwise Student’s t-test for the

mean activity. Values smaller than the significance level (α

= 0.05) are presented in italic.

-value S1 S2 S3 S4 S5

S2 < 0.001

S3 < 0.001 0.536

S4 0.01

0.083 0.112

S5 0.105 0.259 0.356 0.374

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

214

3.3 Performance Throughout the

Game

Similar to the previous analyses, the performance of

the athletes was studied with respect to the different

sets. Figure 3 indicates the mean value and standard

deviation for the total number of actions performed

per athlete. Figure 4 shows the mean value and

standard deviation for the number of minutes during

which the individuals were actively participating in

the game (i.e. not substituted).

Figure 3: Mean and standard deviation for the number of

actions performed during each set.

Figure 4: Mean and standard deviation for the time played

during each set.

For both performance variables, no significant

differences were observed across the five sets

according to the Kruskal-Wallis test. A p-value of

0.052 and 0.121 was obtained for the data in Figure 3

and Figure 4 respectively.

4 DISCUSSION

First, we note that the mean value for the heart rate

decreases significantly from set 1 to 4 in our test

population. On average, a decrease of 5.2 %HR

max

(or

10.5 bpm in absolute numbers) is observed. The

decreasing trend in the heart rate can partially be

explained by the decrease in the total activity

obtained from the accelerometer data. However, the

differences in the activity are only significant for set

1 compared to all other sets.

The decrease in %HR

max

can be interpreted as a

decrease in exercise intensity. In literature, González

et al. (2005) also noted a decrease in the heart rate

over the different sets. These observations might be

linked to fatigue. Initially, athletes play the game at a

high level of exercise intensity, but as the competition

continues, the exercise intensity gradually decreases

because athletes are getting tired. With a total game

duration between about 70 to 120 minutes, this seems

to be a valid hypothesis. However, no reference

measure for the level of fatigue was used in this study.

Additionally, we also note that fatigue is often linked

to cardiovascular drift, which is a gradual intensity-

independent increase in heart rate (Achten &

Jeukendrup, 2003). While these data show an

opposite trend, they also show a decrease in the

external load (i.e. activity). Therefore, the analogy

with cardiac drift is difficult to make.

Other factors such as stress and game tactics are

also likely to affect the measured variables (Schneider

et al., 2018). Stress was not considered in this study.

The performance of the athlete for the different sets

could indicate a change in the tactics. However, both

performance variables did not seem to vary. One

would expect there to be a difference between set 5

and all other sets since the fifth set is ended at a score

of 15 instead of 25. From the visual representation of

the data, these expected differences seem to be

present. However, the limited amount of observations

for the fifth set limits the statistical power of this

analysis.

We have to bear in mind that the analyses are

based on data of the entire test population. If we

consider the data for each individual, the decreasing

trend in the heart rate is more pronounced in certain

athletes, whereas the trend is barely or not significant

in others. Also, the current population only consists

of eight athletes. For the generalisation of these

observations, data collection has to be performed on

a larger scale.

5 CONCLUSIONS

Surprisingly, very few studies were found on the heart

rate and exercise intensity of volleyball players

during competition. Accordingly, a comparison of

our data with other observations was not in order.

This research illustrates how the heart rate (%HR

max

activity and performance change over the course of a

Heart Rate and Activity Measured during Volleyball Competition using Wearable Technology

215

volleyball game. Although each match is unique, a

significant decreasing trend in the heart rate is

observed on the level of the population. Besides, only

a significantly higher activity is observed in the first

set. Finally, the performance of the athlete (expressed

in the number of actions and minutes played) does not

show a significant trend over the course of the set.

Monitoring volleyball athletes using wearable

technology, during practice and competition, can help

to understand the exercise intensity and requirements

of the athletes. This research provides a first overview

of how the heart rate, activity and performance

change during competition. More detailed analyses,

for example on the dynamics of the heart rate in

relation to the recovery periods, will lead to new

insights and provides us with tools to assess the status

of the athlete during each phase of the competition.

Real-time information on the athlete’s physiological

status could provide coaches and staff with additional

information to make tactical decisions during the

game.

For future research, we advise using reference

measures for factors such as stress, fatigue, recovery,

etc. and to adopt a more advanced time series analysis

approach. Furthermore, a larger test population will

lead to generalisation of the data and identifying

potential differences between positions on the field.

ACKNOWLEDGEMENTS

The authors would first like to thank the athletes who

participated in this study. Collecting data during

competition is not self-evident, since it requires effort

and does not directly contribute to better

performance. We would like to thank the coaches and

staff of the teams as well as Sport Vlaanderen for

collaborating in this study.

REFERENCES

Achten J., Jeukendrup A. E., 2003, Heart rate monitoring:

applications and limitations. Sports Medicine, 33, pp

517-538

Arts F. J., Kuipers H., 1994, The relation between power

output, oxygen uptake and heart rate in male athletes.

International Journal of Sports Medicine, 15, pp 228-

231

Buchheit M., 2014. Monitoring training status with HR

measures: Do all roads lead to Rome?, Frontiers in

Physiology, 60, pp 1-19.

Gionet N., 1980. Is volleyball an aerobic or anaerobic

sport?, Volleyball technical journal, 5, pp 31-36.

González C., Ureña A., Llop F., García J. M., Martín A.,

Navarro F., 2005, Physiological characteristics of libero

and central volleyball players. Biology of Sport, 22.

Jimenez-Olmedo J. M., Pueo B., Penichet-Tomás A.,

Chinchilla-Mira J. J., Perez-Turpin J. A., 2017,

Physiological work areas in professional beach

volleyball: A case study, Retos, 31, pp 94-97.

Künstlinger U., Ludwig H. G., Stegemann J., 1987,

Metabolic Changes During Volleyball Matches,

International Journal of Sports Medicine, 8, pp 315-

322.

Schneider C., Hanakam F., Wiewelhove T., Döweling A.,

Kellmann M., Meyer T., Pfeiffer M., Ferrauti A., 2018,

Heart Rate Monitoring in Team Sports - A Conceptual

Framework for Contextualizing Heart Rate Measures

for Training and Recovery Prescription, Frontiers in

Physiology, 9, 639

Sheppard J., Gabbett T., Riggs M., 2013. ‘Indoor and Beach

Volleyball Players’ In Tanner R. K. (ed.), Physiological

testing of elite athletes, 2

edition, Human Kinetics, pp

475-486.

Van Heest J. L., 2003 ‘Energy demands in the sport of

volleyball’, in Reese J. C. (ed.), Handbook of Sports

Medicine and Science: Volleyball, Blackwell

Publishing, pp 11-17.

Hopkins W. G., 1991. Quantification of training in

competitive sports: methods and applications. Sports

Medicine, 12, pp 161-183

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

216