Validity and Reliability of the Myotest Accelerometric System for the Assessment of Vertical Jump Height

Abstract

Casartelli, N, Müller, R, and Maffiuletti, NA. Validity and reliability of the Myotest accelerometric system for the assessment of vertical jump height. J Strength Cond Res 24(11): 3186-3193, 2010-The aim of the present study was to verify the validity and reliability of the Myotest accelerometric system (Myotest SA, Sion, Switzerland) for the assessment of vertical jump height. Forty-four male basketball players (age range: 9-25 years) performed series of squat, countermovement and repeated jumps during 2 identical test sessions separated by 2-15 days. Flight height was simultaneously quantified with the Myotest system and validated photoelectric cells (Optojump). Two calculation methods were used to estimate the jump height from Myotest recordings: flight time (Myotest-T) and vertical takeoff velocity (Myotest-V). Concurrent validity was investigated comparing Myotest-T and Myotest-V to the criterion method (Optojump), and test-retest reliability was also examined. As regards validity, Myotest-T overestimated jumping height compared to Optojump (p < 0.001) with a systematic bias of approximately 7 cm, even though random errors were low (2.7 cm) and intraclass correlation coefficients (ICCs) where high (>0.98), that is, excellent validity. Myotest-V overestimated jumping height compared to Optojump (p < 0.001), with high random errors (>12 cm), high limits of agreement ratios (>36%), and low ICCs (<0.75), that is, poor validity. As regards reliability, Myotest-T showed high ICCs (range: 0.92-0.96), whereas Myotest-V showed low ICCs (range: 0.56-0.89), and high random errors (>9 cm). In conclusion, Myotest-T is a valid and reliable method for the assessment of vertical jump height, and its use is legitimate for field-based evaluations, whereas Myotest-V is neither valid nor reliable.

Introduction

Vertical jump performance can be assessed with a variety of test procedures, which include different measurement tools (e.g., force platforms, video analysis systems, contact mats, photoelectric cells), jump modalities (e.g., squat jump [SJ], countermovement jump [CMJ], drop jump, repeated jumps [RJs]), and a variety of performance calculation methodologies (e.g., vertical takeoff velocity, flight time, mechanical power, body center of mass displacement). Depending on the measurement device, vertical jump assessment can be classified as laboratory based or field based. The former generally involves the measurement of jump performance with force platforms or video analysis systems. These assessments have excellent measurement accuracy (^{2,5,8,11,13,15,16}), but they require expensive and hardly transportable measurement tools. In contrast, field-based assessments using contact mats or photoelectric cells are easily accessible to physical trainers and sport scientists who want to measure the vertical jump performance in the same conditions (e.g., same surface) as actual sport activity.

The Myotest system (Myotest SA) is a newly developed measurement tool specifically designed for field-based assessment of vertical jump ability. It consists of a single accelerometer contained in a small and foolproof device. The device is attached to the hip level during the execution of a vertical jump so as to record the vertical acceleration of the subject's body. Based on vertical takeoff velocity or flight-time recordings, the Myotest system can then provide an estimation of vertical jump height. Compared to the other devices for field-based jumping evaluation, it has the advantages of being extremely small (5 × 10 cm) and portable, easy to handle, relatively cheap, and also of being used on particular surfaces (e.g., on the sand).

However, the control for the validity and reliability of the Myotest system is a fundamental step for accepting it as an appropriate system for vertical jump height assessment. Validity refers to the ability of the measurement tool to reflect what is designed to measure. Reliability can be defined as the consistency of measurements. If the Myotest system will show good concurrent validity, it could replace an already validated measurement tool. Moreover, test-retest reliability analyses will ascertain the absence of measurement error. Validity and reliability of Myotest measurements have been verified for dynamic variable-resistance weight lifting, with the device applied to a barbell during the bench press (¹²). Myotest recordings resulted in being both valid and reliable up to 70% of the 1-repetition maximum; however, beyond this load, data were no longer valid and reliable, because of the slow velocity of execution. To our knowledge, the validity and reliability of the Myotest system for the assessment of vertical jump height have never been examined, despite large use of this device in elite and subelite athletes, particularly in team sports (see www.myotest.ch).

Therefore, the aim of the present study was to test the validity and reliability of the Myotest accelerometric system for the assessment of single and repeated vertical jump heights in basketball players of different ages. If the Myotest system will result both valid and reliable for jumping height measurement, its use will be legitimate for field-based assessments.

Methods

Experimental Approach to the Problem

Two identical test sessions separated by 2-15 days were organized. Vertical jump flight height, that is, the difference between the height of the subject's body center of mass at the peak of the jump and at takeoff (^9,14), was simultaneously quantified with the Myotest system and with ground-based photoelectric cells (Optojump, Microgate, Bolzano, Italy). The latter measurement system has been recently validated for vertical jump height assessment in our laboratory by comparing it to a force platform (¹⁰), which is usually considered the gold standard. Therefore, we adopted the ground-based photoelectric cells as the criterion instrument in this study, because it is more practical to use in the field than a force platform. Two calculation methods were used to estimate the jump height from Myotest recordings: flight time (Myotest-T) and vertical takeoff velocity (Myotest-V). Concurrent validity was investigated comparing the Myotest-T and Myotest-V data to the criterion method (Optojump), and test-retest reliability of each method was also examined. As a methodological check, video analysis, which is assumed to be a very accurate method for assessing vertical jump performance (⁵), was also implemented. We tested basketball players, because vertical jump is 1 of the most prevalent activities performed in this sport (¹⁷), and we considered 3 of the most used jump modalities in basketball practice and testing (¹⁷): SJ, CMJ, and RJs.

Subjects

Forty-four healthy male basketball players with a mean (±SD) age of 15.3 ± 3.8 years (range: 9-25 years), height of 178 ± 18 cm (range: 141-210 cm), and body mass of 68 ± 18 kg (range: 30-101 kg) volunteered to participate in this study. All players belonged to the same basketball club but to different age-group teams. They all competed in regional and national Swiss basketball championships. Their training volume ranged from 4 h·wk⁻¹ for the youngest players to 12 h·wk⁻¹ for the oldest ones. Subjects and parents of minor subjects were informed about the procedures of the study, and signed a written informed consent, which was approved by the ethics committee of the Swiss Federal Institute of Technology of Zurich.

Procedures

Testing

For both test sessions, experimental conditions (e.g., time of day, fatigue, sleep, motivation) were strictly controlled by means of a questionnaire, which was completed by the experimenter (N.C.) at the end of each test. Before each test session, subjects performed a standardized warm-up routine that consisted of approximately 3 minutes of jogging at a comfortable pace and 2 minutes of static stretching of the lower extremity muscles. They were also instructed on how to perform the different jumps and completed familiarization trials until the correct technique was learned (as judged by the experimenter). For SJ, subjects started from the upright standing position with their hands on their hips; they were then instructed to flex their knees and hold a predetermined knee position (ca. 90°), and the experimenter then counter out for 3 seconds. On the count of 3 subjects were instructed to jump as high as possible without performing any countermovement before the execution of the jump. For CMJ, subjects started from the upright standing position with their hands on their hips; they were then instructed to flex their knees (ca. 90°) as quickly as possible and then jump as high as possible in the ensuing concentric phase. For RJs, subjects were instructed to start with a CMJ, and as soon as the feet touched the ground at landing they were asked to jump up again as high as possible keeping the knees and ankles extended to minimize the contact with the ground. With all jumps, it was recommended that at takeoff, the subjects leave the floor with the knees and ankles extended and land in a similarly extended position. If a jump was incorrectly performed, as judged by the experimenter, subjects were asked to repeat the trial. A total of 15 maximal vertical jumps were evaluated, which included 5 SJs, 5 CMJs, and 5 consecutive RJs. The rest interval between each SJ and CMJ was approximately 30 seconds. The rest interval between series of jumps was approximately 3 minutes.

Vertical Jump Height Estimation

The Myotest device (dimensions: 5.4 × 10.2 × 11.1 cm; weight: 58 g) contains a 3D inertial accelerometer (±8 g), which allows vertical acceleration to be recorded at a sampling frequency of 500 Hz. The device was perpendicularly attached to a large (8.5 cm) Velcro elastic belt provided with the Myotest device. The device was fixed to the hip level on the left side of the body, as indicated by the manufacturer (Figure 1).

Figure 1: The Myotest device presented beside a 1 Swiss franc coin (A), attached to the proper Velcro elastic belt (B), and fixed to the hip level on the left side of the body for accelerometric recordings during vertical jump (C).

Accelerometric data were stored during the assessments and subsequently downloaded for jump height calculations using the proprietary software (Myotest PRO Software version 1.0). The software automatically integrated the acceleration recordings to obtain vertical velocity, from which jump height was estimated using 2 different calculation methods. The first method (Myotest-T) used the flight time (i.e., the time interval between the peak positive and the peak negative vertical velocity), according to the following equation: (jump height = [g × flight time²]/8) (⁶). The second method (Myotest-V) used the highest vertical takeoff velocity, according to the equation: (jump height = maximal vertical velocity²/[2 × g]) (¹⁴).

The 2 Optojump bars (length: 1 m) were placed parallel on the ground at a distance of 1 m to each other (Figure 1). Flight-time data were recorded with a sampling frequency of 1,000 Hz, and Optojump vertical jump height was automatically calculated by the proprietary software (Optojump version 3.01.0001) using the same equation as Myotest-T (⁶). These test procedures have been shown to result in a valid and reliable assessment of vertical jump height, when compared to a force plate (¹⁰).

A commercially available video camera (JVC GR-DVL 167, Wayne, NY, USA) with a sampling frequency of 50 Hz and a shutter speed of 1/120 seconds was used to monitor all vertical jumps. A marker was attached to the left lateral femoral condyle, and its displacement from the starting position to peak jump height was quantified using Peak Motus System version 9.2 (Vicon Motion Systems, Oxford, United Kingdom).

Statistical Analyses

For the different jump modalities (SJ, CMJ, and RJ) and measurement methods (Myotest-T, Myotest-V, Optojump), the mean value of the 5 trials was retained. Shapiro-Wilk tests revealed that all jump heights were normally distributed. One-way ANOVAs with repeated measures were used to detect any systematic bias between Myotest (Myotest-T, Myotest-V) and Optojump data. Post hoc analyses (Tukey's honestly significant difference [HSD] test) were then used to test for differences among pairs of means. Concurrent validity was assessed using the intraclass correlation coefficient (ICC) (^2,1) and Bland-Altman method systematic bias ± random error (³). Because of the presence of heteroscedasticity, that is, absolute differences related (R² > 0.1) to the magnitude of the mean (¹), limits of agreement (LOA) ratios were also calculated (⁷). Test-retest reliability was assessed with ICC (^2,1) (relative reliability), coefficients of variation (CVs), and Bland-Altman method (³) (absolute reliability). Paired Student t-tests were used to detect any systematic bias between test and retest results. For all statistical procedures, we used the Bonferroni correction to maintain the overall probability of obtaining a type I error at 0.05 (⁴). Thus, because there are 3 primary outcome variables, a corrected alpha level of 0.017 was accepted as significant.

Results

Concurrent Validity

Regardless of the jump modality (SJ, CMJ, and RJ), greater jumping heights were observed for both Myotest-T (27%) and Myotest-V (34%) compared to Optojump (p < 0.001) (Figure 2).

Figure 2: Jump heights of squat jump (SJ; upper panel), countermovement jump (CMJ; middle panel), and repeated jumps (RJs; lower panel) measured with Myotest-T, Myotest-V, and Optojump. The lower margin of the vertical box indicates the 25th percentile, the line in the middle of the box the median, and the upper margin the 75th percentile of the recorded jump heights. The lower and upper error bars indicate the 10th and 90th percentiles, respectively. *Myotest-T and Myotest-V jump heights are significantly higher than those of Optojump at the p < 0.001.

Bland-Altman plots (Figure 3) revealed an average systematic bias of 7.2 cm (p < 0.001) between Myotest-T and Optojump data (Table 1). For the comparison between Myotest-V and Optojump (Figure 3, Table 1), systematic bias ranged between 5.7 and 12.3 cm (p < 0.001), but CMJ and RJ showed heteroscedasticity, that is, the difference between the 2 methods increased with increasing jumping height. Regardless of the jump modality, random errors and LOA ratios were considerably higher for Myotest-V than for Myotest-T (Table 1).

Table 1: Concurrent validity of Myotest vs. Optojump data.*

Figure 3: Validity: Bland-Altman plots of Myotest-T vs. Optojump (left) and Myotest-V vs. Optojump (right) jump heights for squat jump (SJ; upper panels), countermovement jump (CMJ; middle panels), and repeated jumps (RJ; lower panels). The short-dashed lines represent the upper and lower 95% limits of agreement, whereas the solid line represents the bias. The long-dashed line represents the correlation between differences and average values.

Similarly, ICCs for concurrent validity were higher (>0.98) for Myotest-T than for Myotest-V (<0.75) (Table 1).

Test-Retest Reliability

Test-retest ICCs were high and comparable between Myotest-T (range: 0.92-0.96) and Optojump (range: 0.93-0.97), whereas Myotest-V showed lower and insufficient ICCs (range: 0.56-0.89) (Table 2). In the same way, both random errors and CVs were relatively low and similar for Myotest-T and Optojump, whereas they were 2-3 times greater for Myotest-V, RJ in particular. A significant test-retest systematic bias was observed for SJ, as assessed with all 3 methods, and for CMJ height, as assessed using Myotest-T, with lower jumping heights in the retest condition.

Table 2: Test-retest reliability of Myotest and Optojump data.*

Discussion

Myotest-T showed excellent concurrent validity and test-retest reliability for the assessment of single and repeated jumping height, despite significant overestimation compared to the criterion instrument Optojump (¹⁰). In contrast, Myotest-V showed poor concurrent validity and insufficient reliability.

The systematic overestimation of jump height using Myotest-T compared to Optojump was probably caused by the jump height calculation methodology. Theoretically, the flight time of a vertical jump corresponds to the time elapsed from the instant the subject's feet leave the ground (i.e., takeoff) to ground contact (i.e., landing in the same position). The most accurate tools to record vertical jump flight time are force platforms or ground-based photoelectric cells (¹⁰), which allow takeoff-landing moments to be precisely identified. Myotest-T estimates vertical jump height using the time difference between the positive (propulsive phase of the jump) and negative (landing) peaks of vertical velocity. However, the maximal positive velocity is normally achieved shortly before the takeoff, and the maximal negative velocity shortly after landing, which inevitably causes the flight time recorded by Myotest-T to be longer than the effective flight time. It must also be considered that photoelectric cells systematically underestimate vertical jump height by approximately 1 cm when compared to a force platform (¹⁰), which would contribute to increase the intermethod differences reported here. Glatthorn et al. (¹⁰) demonstrated that the position of the cells, which are elevated from the ground by few millimeters, causes the recorded flight time and therefore the jump height to be slightly underestimated compared to a force plate. Taken as a whole, these results suggest that the jump height difference between Myotest-T and force platform would be approximately 6 cm. In the same way, the calculation methodology seems responsible, at least partly, for the overestimation of vertical jump height using Myotest-V. Myotest-V employs the vertical takeoff velocity method to calculate the jump height (¹⁴). However, the Myotest system uses the positive peak of vertical velocity and not the real takeoff velocity. As discussed above, during a vertical jump, the maximal positive velocity is usually reached slightly before the takeoff, so that takeoff velocity (and therefore jump height) is necessarily lower than maximal velocity. As a consequence, greater vertical jump height is obtained with Myotest-V compared to the criterion instrument.

Although Myotest-T and Myotest-V both use the same vertical velocity recordings and both apply incorrect calculations to estimate the height of jump, Myotest-T showed excellent concurrent validity, whereas Myotest-V showed poor validity. Because the former employs the flight time to estimate the vertical jump height (⁶), the absolute vertical velocity recorded by the Myotest is not directly involved in calculations. On the other hand, Myotest-V uses the maximal positive velocity to estimate vertical jump height (¹⁴). As a possible interpretation, we believe that the Myotest system can successfully detect the instants of takeoff (at peak positive velocity) and landing (at peak negative velocity), whereas it incorrectly quantifies the vertical velocity of the subject's body during maximal vertical jumping. It is indeed possible that the Myotest device undergoes a slight displacement as a function of the subject's body during the execution of single and RJs. Despite clearly established procedures for jump execution and Myotest fixation, between-subject differences in (a) jumping strategy, (b) elastic belt fastening and positioning around the hips, and (³) Myotest device orientation as a function of the elastic belt could account for unexpected device displacements during jump execution, which would subsequently affect accelerometric-based determination of vertical velocity.

Despite absolute differences in vertical jump heights between video analysis and the other calculation methods (data not reported), because actual measurement of jump height with video analysis does not correspond to the estimation of flight height (⁹), we observed high ICCs between Myotest-T and video analysis (range: 0.94-0.96) that further confirm the validity of the former calculation method. Video analysis also corroborated the poor concurrent validity of Myotest-V, as witnessed by low ICC values (range: 0.56-0.79).

Although a systematic bias of approximately 1 cm was observed between test and retest for Myotest-T, its relative and absolute reliability scores were not compromised. Both systematic biases and random errors were indeed very similar to those reported for the criterion instrument. It seems that the slightly lower jump heights obtained at retest were caused by a general lower motivation level in comparison to the first assessment. During the first test, subjects were strongly motivated by the novelty of the jumping test methodology, whereas at retest, they were already familiar with the entire testing process, so therefore, their motivation probably was lower.

In the present study, the sensitivity of Myotest-T calculation method was 99%, that is, it detected correct jump heights for 1,304 of the 1,320 recorded jumps. In the remaining 16 cases, the Myotest analysis software showed jump heights of 0 cm or considerably lower than those provided by the Optojump. We believe Myotest-T incorrectly detected the maximal negative velocity (i.e., the instant of landing) or did not detect it at all. In contrast, the sensitivity of Myotest-V was 100%, that is, maximal vertical velocity could always be recorded.

A limitation of this study is the use of ground-based photoelectric cells, instead of a force platform (gold standard), as the criterion instrument for the validation of the Myotest system. However, the photocells used in this study have been shown to have strong concurrent validity and excellent test-retest reliability in comparison to a force platform (¹⁰) and are more practical to use in the field. Another limitation is that gyroscopes were not used to correct Myotest vertical acceleration recordings to the device inclination angle during jumping. Because the device was applied to an elastic belt, perpendicular to the ground, when the subject was standing before jumping, it is possible that it moved forward a certain amount during the propulsive phase of the jump due to trunk bending. This would have caused vertical acceleration and consequently vertical velocity and flight time recordings to present a certain amount of random error.

Practical Applications

Myotest accelerometric system is a valid and reliable tool for the assessment of vertical jump height on the field, provided Myotest-T (and not Myotest-V) calculation method is used, as also indicated by the manufacturer. Therefore, this device can be used with confidence to detect within-group changes in longitudinal assessments (e.g., to verify the effectiveness of a specific training program, to quantify possible alterations during the competitive season) and between-group differences in cross-sectional comparisons (e.g., for talent detection). Compared to the other devices for field-based jumping evaluation (photoelectric cells and contact mats), Myotest has the advantages of being extremely portable and easy to use, relatively inexpensive, and also to respect the specificity between sport activities and jumping evaluation (e.g., it can be used on sand). However, Myotest results are not interchangeable with respect to other measurement instruments, because of the systematic overestimation of jumping height.

Acknowledgments

The authors gratefully acknowledge Mr. Patrick Flaction (Myotest SA, Sion, Switzerland) for lending us the Myotest system, and Ms. Kirsten Dobson for English proofreading of the manuscript. None of the authors have received any payment, grant, research support, or other financial support or payment related to this work.