Professional Documents
Culture Documents
Journal of Strength and Conditioning Research
Journal of Strength and Conditioning Research
Reliability and concurrent validity of seven commercially available devices for the
assessment of movement velocity at different intensities during the bench press
--Manuscript Draft--
Full Title: Reliability and concurrent validity of seven commercially available devices for the
assessment of movement velocity at different intensities during the bench press
Keywords: linear position transducer; linear velocity transducer; smartphone application, inertial
measurement units; Velocity-based training; testing
Antonio Piepoli
Gabriel Delgado-García
Gabriel Garrido-Blanca
Abstract: This aim of this study was to compare the reliability and validity of seven commercially
available devices to measure movement velocity during the bench press exercise.
Fourteen men completed two testing sessions. The bench press one-repetition
maximum (1RM) was determined in the first session. The second testing session
consisted of performing three repetitions against five loads (45-55-65-75-85% of 1RM).
The mean velocity was simultaneously measured using an optical motion sensing
system (Trio-OptiTrack™; “gold-standard”) and seven commercially available devices:
1 linear velocity transducer (T-Force™), 2 linear position transducers (Chronojump™
and Speed4Lift™), 1 camera-based optoelectronic system (Velowin™), 1 smartphone
application (PowerLift™), and 2 inertial measurement units (PUSH™ band and
Beast™ sensor). The devices were ranked from the most to the least reliable as
follows: (I) Speed4Lift™ (coefficient of variation [CV] = 2.61%), (II) Velowin™ (CV =
3.99%), PowerLift™ (3.97%), Trio-OptiTrack™ (CV = 4.04%), T-Force™ (CV = 4.35%),
Chronojump™ (CV = 4.53%), (III) PUSH™ band (CV = 9.34%), and (IV) Beast™
sensor (CV = 35.0%). A practically perfect association between the Trio-OptiTrack™
system and the different devices was observed (Pearson’s product-moment correlation
coefficient (r) range = 0.947-0.995; P < 0.001) with the only exception of the Beast
sensor (r = 0.765; P < 0.001). These results suggest that linear velocity/position
transducers, camera-based optoelectronic systems and the smartphone application
could be used to obtain accurate velocity measurements for restricted linear
movements, while the inertial measurement units used in this study were less reliable
and valid.
Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Title Page (Showing Author Information)
Reliability and concurrent validity of seven commercially available devices for the
Institutional Affiliations:
1
Department of Physical Education and Sport, Faculty of Sport Sciences, University of
Corresponding author:
Acknowledgments:
The authors thank all the participants who selflessly participated in the study. This study was
supported by the Spanish Ministry of Education, Culture and Sport under a pre-doctoral grant
(FPU15/03649) awarded to APC and by the University of Granada under a postdoctoral grant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27 Reliability and concurrent validity of seven commercially available devices for the
28
29 assessment of movement velocity at different intensities during the bench press
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
1
65
Feasibility of seven devices to measure velocity
ABSTRACT
1
2 This aim of this study was to compare the reliability and validity of seven commercially
3
4
5 available devices to measure movement velocity during the bench press exercise. Fourteen
6
7 men completed two testing sessions. The bench press one-repetition maximum (1RM) was
8
9
10 determined in the first session. The second testing session consisted of performing three
11
12 repetitions against five loads (45-55-65-75-85% of 1RM). The mean velocity was
13
14
15 simultaneously measured using an optical motion sensing system (Trio-OptiTrack™; “gold-
16
17 standard”) and seven commercially available devices: 1 linear velocity transducer (T-
18
19
Force™), 2 linear position transducers (Chronojump™ and Speed4Lift™), 1 camera-based
20
21
22 optoelectronic system (Velowin™), 1 smartphone application (PowerLift™), and 2 inertial
23
24 measurement units (PUSH™ band and Beast™ sensor). The devices were ranked from the
25
26
27 most to the least reliable as follows: (I) Speed4Lift™ (coefficient of variation [CV] =
28
29 2.61%), (II) Velowin™ (CV = 3.99%), PowerLift™ (3.97%), Trio-OptiTrack™ (CV =
30
31
32 4.04%), T-Force™ (CV = 4.35%), Chronojump™ (CV = 4.53%), (III) PUSH™ band (CV =
33
34 9.34%), and (IV) Beast™ sensor (CV = 35.0%). A practically perfect association between the
35
36
37
Trio-OptiTrack™ system and the different devices was observed (Pearson’s product-moment
38
39 correlation coefficient (r) range = 0.947-0.995; P < 0.001) with the only exception of the
40
41 Beast sensor (r = 0.765; P < 0.001). These results suggest that linear velocity/position
42
43
44 transducers, camera-based optoelectronic systems and the smartphone application could be
45
46 used to obtain accurate velocity measurements for restricted linear movements, while the
47
48
49 inertial measurement units used in this study were less reliable and valid.
50
51
52
53
54 Keywords: linear position transducer; linear velocity transducer; smartphone application,
55
56 inertial measurement units; velocity-based training; testing.
57
58
59
60
61
62
63
64
2
65
Feasibility of seven devices to measure velocity
INTRODUCTION
1
2 Velocity-based resistance training has gained in popularity over recent years due to the
3
4
5 proliferation of different commercially available devices (e.g., linear position transducers,
6
7 inertial measurement units [IMUs], smartphone applications, etc.) that are supposed to
8
9
10 accurately measure movement velocity (3,5). It has been proposed that the monitoring of
11
12 barbell velocity could be an appropriate alternative to prescribe the training load as compared
13
14
15 to the traditional approach that requires the determination of the one-repetition maximum
16
17 (1RM) (16,32). The use of movement velocity to prescribe the training load is justified by the
18
19
strong and linear relationship that has been reported for multiple exercises between
20
21
22 movement velocity and the %1RM (12,26,31). In this regard, instead of determining the 1RM
23
24 through a single maximal lift or by a set of repetitions to failure, the load can be prescribed to
25
26
27 match the desired velocity (16,34). Despite the encouraging applications of velocity-based
28
29 resistance training (20), little research is available comparing the reliability and validity of
30
31
32 different commercially-available devices used in training and research to monitor movement
33
34 velocity.
35
36
37
38
39 From a scientific standpoint, the three-dimensional (3D) motion capture has been
40
41 recognized as the “gold-standard” instrument to measure movement velocity (22,36).
42
43
44 However, since this technology is not practical or affordable for strength and conditioning
45
46 professionals, others devices are typically used in practice when implementing the velocity-
47
48
49 based resistance training approach. The linear position transducer has been the most used
50
51 device in scientific research (2,5,9,13). The linear position transducer consists of an
52
53
54 isoinertial dynamometer with a cable that is typically attached to the barbell and it derives
55
56 velocity from the recorded displacement-time data using the inverse dynamic approach (17).
57
58
More recently, a linear velocity transducer named “T-force™” (T-Force™ system, Ergotech,
59
60
61
62
63
64
3
65
Feasibility of seven devices to measure velocity
Murcia, Spain) has been made commercially available, which directly provides velocity
1
2 measurements through the recording of electrical signals that are proportional to the cable’s
3
4
5 extension velocity (33). It is reasonable to speculate that the linear velocity transducer could
6
7 be more precise than linear position transducers since it is known that the successive
8
9
10 manipulation of raw data increases measurement errors (25,30). However, it remains
11
12 unexplored whether the reliability of velocity outputs significantly differs between linear
13
14
15 position and linear velocity transducers, as well as their concurrent validity with respect to
16
17 the “gold standard” 3D motion capture.
18
19
20
21
22 It should be acknowledged that linear position/velocity transducers are not always
23
24 practical or affordable. The need to attach the cable to the barbell restricts exercise selection
25
26
27 to the ones predominantly performed in a vertical direction (5). Another common drawback
28
29 of linear position/velocity transducers is their high price ( 2,000 US dollars), which may
30
31
32 limit their use to laboratory-based or professional sport settings (6,23,24). However, it should
33
34 be noted that a new linear position transducer named “Speed4Lift™” (Speed4Lift™, Madrid,
35
36
37 Spain) have appeared on the market with a considerably lower price (340 US dollars),
38
39 although there are no available data regarding its reliability and validity. As an alternative to
40
41
42
linear position/velocity transducers, wearable technologies are increasingly gaining
43
44 popularity in the field of strength training and conditioning (3,5,11,28).
45
46
47
48
49 One of the wearable devices that have recently appeared on the market is named
50
51 “Velowin™” (Velowin™, DeporTeC, Murcia, Spain). Velowin™ is a camera-based
52
53
54 optoelectronic system designed to measure movement velocity by the tracking of an infrared
55
56 reflective marker placed in the barbell. A high reliability and concurrent validity of the
57
58
59 Velowin™ to measure movement velocity has been reported during the free-weight back
60
61
62
63
64
4
65
Feasibility of seven devices to measure velocity
To address the existing gaps in the literature, the present study was designed to
1
2 provide a comprehensive analysis of different devices (i.e., linear velocity transducer, linear
3
4
5 position transducers, camera-based optoelectronic system, smartphone application and IMUs)
6
7 that are being used in practice for the measurement of movement velocity during resistance
8
9
10 training. Specifically, the objective of this study was to compare the reliability and validity of
11
12 of seven commercially available devices to measure movement velocity during the bench
13
14
15 press exercise. We hypothesized that the devices would be ranked from the most to the least
16
17 reliable and valid as follows: (I) linear velocity transducer; (II) linear positions transducers;
18
19
(III) camera-based optoelectronic device, (IV) smartphone application and (V) IMUs. The
20
21
22 results of this study should provide practical information for strength and conditioning
23
24 coaches regarding the reliability and concurrent validity of different devices that can be used
25
26
27 in practice for the assessment of movement velocity.
28
29
30
31
32 METHODS
33
34 Experimental approach to the problem
35
36
37
This study was designed to explore the reliability and concurrent validity of seven
38
39 commercially available devices for the measurement of movement velocity. Subjects
40
41 completed two testing sessions separated by 48-72 hours. The 1RM in the bench press
42
43
44 exercise was determined in the first testing session. The second testing session consisted of
45
46 performing three repetitions against five different loads (45%, 55%, 65%, 75%, and 85% of
47
48
49 1RM). The mean velocity of the barbell was measured using an optical motion sensing
50
51 system (V120:Trio, OptiTrack™, NaturalPoint Inc., USA) that was considered the gold-
52
53
54 standard in the present study (27,38). In addition, the mean velocity was also measured by
55
56 seven commercially available devices: 1 linear velocity transducer (T-Force™ system,
57
58
Ergotech, Murcia, Spain), 2 linear position transducers (Chronojump™ Boscosystem,
59
60
61
62
63
64
6
65
Feasibility of seven devices to measure velocity
Thereafter, the external load was incremented from 10 to 1 kg until the 1RM load was
1
2 reached. The average number of loads tested was 8.9 ± 1.3. The inter-set rest was set to 4 min
3
4
5 and 1-2 repetitions were performed with each load.
6
7
8
9
10 The second testing session began with the same warm-up described for session 1.
11
12 Afterwards, subjects performed the concentric-only bench press exercise against five relative
13
14
15 loads (45%, 55%, 65%, 75%, and 85% of 1RM) that were implemented in an incremental
16
17 order. Lower loads (i.e., < 45% of 1RM) were not tested because they are generally not used
18
19
in training with the bench press exercise, while higher loads (i.e., > 85% of 1RM) were
20
21
22 excluded to avoid high fatigue that could compromise reliability analyses. Three repetitions
23
24 were executed with each load. Inter-repetition rest was set to 15 s and inter-set rest was fixed
25
26
27 to 4 min. The barbell was held by the safety stops of the Smith machine during the recovery
28
29 periods. Subjects were encouraged to lift the barbell at the maximum possible velocity.
30
31
32
33
34 The standard five-point body contact position technique (head, upper back, and
35
36
37
buttocks firmly on the bench with both feet flat on the floor) was followed in the two testing
38
39 sessions. Subjects self-selected the grip width, which was measured and kept constant on
40
41 every lift. The barbell was held by the safety stops of the Smith machine 1-2 cm above
42
43
44 subjects' chest at the level of the sternum. From that position, subjects were instructed to
45
46 perform a concentric-only movement as fast as possible until their elbows reached full
47
48
49 extension. Two spotters were responsible for lowering the barbell after each repetition.
50
51
52
53
54 Measurement equipment and data acquisition
55
56
57
58
59
60
61
62
63
64
8
65
Feasibility of seven devices to measure velocity
The measurement devices were not synchronized and they separately collected the mean
1
2 velocity of the same repetitions (Figure 1). The specific characteristics of each device are
3
4
5 provided below:
6
7
8
9
10 **Figure 1 near here**
11
12
13
14
15 -Trio-OptiTrack™ (V120:Trio, OptiTrack™, NaturalPoint Inc., USA): an optical
16
17 motion sensing system which included three infrared and precalibrated cameras fixed on a
18
19
rectangular frame that give 3D position data of a reflective marker at a sampling rate of 120
20
21
22 Hz. Raw data of marker position in space were acquired using the software Motive v.1.5.0
23
24 (OptiTrack, NaturalPoint Inc., USA) and then analysed in Microsoft Excel (Microsoft,
25
26
27 Seattle, WA, USA). Instantaneous velocity was calculated by the differentiation of the
28
29 displacement data with respect to time. The reflective marker was placed on the left side of
30
31
32 the barbell and the Trio-OptiTrack™ was positioned at a distance of 2.5 m from the marker.
33
34 -T-Force™ (T-Force™ system, Ergotech, Murcia, Spain): an isoinertial dynamometer
35
36
37
which consist of a cable-extension linear velocity transducer interfaced with a personal
38
39 computer by means of a 14-bit resolution analog-to-digital data acquisition board.
40
41 Instantaneous velocity was automatically calculated at a sampling rate of 1,000 Hz by the
42
43
44 custom software v.2.28. The cable was vertically attached to the right side of the barbell
45
46 using a Velcro strap.
47
48
49 -Chronojump™ (Chronojump™ Boscosystem, Barcelona, Spain): an isoinertial
50
51 dynamometer which consist of a cable-extension linear position transducer attached to the
52
53
54 barbell interfaced with a personal computer at a sampling rate of 1,000 Hz. Raw data were
55
56 exported from custom software v.1.6.2 and then analysed in Microsoft Excel (Microsoft,
57
58
Seattle, WA, USA). Instantaneous velocity was calculated by the differentiation of the
59
60
61
62
63
64
9
65
Feasibility of seven devices to measure velocity
displacement data with respect to time. The cable was vertically attached to the right side of
1
2 the barbell using a Velcro strap.
3
4
5 -Speed4Lift™ (Speed4Lift™, Madrid, Spain): an isoinertial dynamometer which
6
7 consist of a cable-extension linear position transducer attached to the barbell. Data were
8
9
10 directly recorded by the differentiation of the displacement data with respect to time at a
11
12 sampling rate of 1,000 Hz via wi-FI connection with an Android smartphone using
13
14
15 Speed4Lift application v.4.1. The cable was vertically attached to the left side of the barbell
16
17 using a Velcro strap.
18
19
-Velowin™ (Velowin™, DeporTeC, Murcia, Spain): an optoelectronic system which
20
21
22 included an infrared camera interfaced with a personal computer that measured displacement
23
24 of a reflector fixed to the barbell. Data were directly recorded from the custom software
25
26
27 v.1.6.314 by the differentiation of the displacement data with respect to time at a sampling
28
29 rate of 500 Hz. The Velowin was placed at a distance of 1.7 m from the infrared reflector and
30
31
32 it was calibrated according to the manufacturer’s instructions.
33
34 -PowerLift™: a smartphone v.6.0.1 application which involves the frame-by-frame
35
36
37
manual inspection of a slow motion video recording by the smartphone high-speed camera at
38
39 a frequency of 240 frames per second and a quality of 720 pixels. The mean velocity was
40
41 computed as the individual range of motion (i.e., vertical displacement of the barbell from the
42
43
44 initial [≈ 1 cm above the subject's chest] to the final [elbows at full extension] position)
45
46 divided by the lifting time (i.e., time between two frames selected by the user). The
47
48
49 smartphone (iPhone, Apple Inc., California, USA) was held by a researcher in portrait
50
51 position and recorded each lift from the front of the subject at approximately 1.5 m.
52
53
54 -PUSH™ band (PUSH™ band, PUSH Inc., Toronto, Canada): a wearable wireless
55
56 IMUs which consist of a 3-axis accelerometer and a gyroscope that provided 6 degrees of
57
58
freedom in its coordinate system (2). Data were directly recorded by the integration of the
59
60
61
62
63
64
10
65
Feasibility of seven devices to measure velocity
acceleration data with respect to time at a sampling rate of 200 Hz via Bluetooth 4.0 LE
1
2 connection with a smartphone (iPhone, Apple Inc., California, USA) using PUSH™
3
4
5 application v.1.1.26. The PUSH™ band was worn on the subject’s dominant forearm
6
7 immediately inferior to the elbow crease with the main button located proximally (2,5).
8
9
10 -Beast™ sensor (Beast™ sensor, Beast Technologies Srl., Brescia, Italy): a wearable
11
12 wireless IMUs which included a 3-axis accelerometer, gyroscope and magnetometer. Data
13
14
15 were directly recorded by the integration of the vertical acceleration with respect to time at a
16
17 sampling rate of 50 Hz via Bluetooth 4.0 LE connection with a smartphone (iPhone, Apple
18
19
Inc., California, USA) using Beast™ application v.2.3.7. The Beast™ sensor was placed on
20
21
22 the barbell using a built-in magnet (3).
23
24
25
26
27 Statistical analyses
28
29 Descriptive data are presented as means and standard deviations, while the coefficient of
30
31
32 variation (CV) of the five loads is presented through their median value. Reliability was
33
34 assessed for each individual load by the CV and the intraclass correlation coefficient (ICC
35
36
37
model 3,1). Acceptable reliability was determined as a CV < 10% and a ICC > 0.70 (7). The
38
39 median CV value of the five loads was calculated to compare the reliability between the eight
40
41 devices examined in the present study. To interpret the magnitude of differences observed
42
43
44 between two CVs, a criterion for the smallest important ratio was established as higher than
45
46 1.15 (37). Bland-Altman plots were constructed to explore the concurrent validity of the
47
48
49 seven commercially available devices respect to the Trio-OptiTrack™ system. Since we
50
51 observed proportional bias in six out of seven comparisons (r2 > 0.1) (1), the data were log-
52
53
54 transformed before calculating the Pearson’s product-moment correlation coefficients (r)
55
56 (18). The criteria to interpret the strength of the r coefficients was as follows: trivial (< 0.1),
57
58
small (0.1-0.3), moderate (0.3-0.5), high (0.5-0.7), very high (0.7-0.9), or practically perfect
59
60
61
62
63
64
11
65
Feasibility of seven devices to measure velocity
(> 0.9) (18). Statistical significance was accepted at P < 0.05 level and confidence limits
1
2 were set at 95%. All reliability assessments were performed by means of a custom
3
4
5 spreadsheet (19), while other statistical analyses were performed using the software package
6
7 SPSS (IBM SPSS version 22.0, Chicago, IL, USA).
8
9
10
11
12 RESULTS
13
14
15 Mean velocity values only reached an acceptable reliability at all loads for the Trio-
16
17 OptiTrack™ system, the linear velocity/position transducers and the smartphone application
18
19
(CV < 6.24% and ICC > 0.73) (Table 1). Based on the comparison of the median CVs of the
20
21
22 5 loads, the devices were ranked from the most to the least reliable as follows: (I)
23
24 Speed4Lift™ (CV = 2.61%), (II) Velowin™ (CV = 3.99%), PowerLift™ (3.97%), Trio-
25
26
27 OptiTrack™ (CV = 4.04%), T-Force™ (CV = 4.35%), Chronojump™ (CV = 4.53%), (III)
28
29 PUSH™ band (CV = 9.34%), and (IV) Beast™ sensor (CV = 35.0%).
30
31
32
33
34 ****Table 1 near here****
35
36
37
38
39 Bland-Altman plots revealed low systematic bias and random errors (≤ 0.05 m·s-1) for
40
41 the T-Force™, Chronojump™, Speed4Lift™, Velowin™, and PowerLift™ as compared to
42
43
44 the Trio-OptiTrack™ system (Figure 2). Both IMUs showed larger random errors (PUSH™
45
46 band = 0.06 ms-1 and Beast™ sensor = 0.21 ms-1). Heteroscedasticity of the errors was
47
48
49 observed for all devices with the only exception of the Speed4Lift™ (r2 = 0.007). A
50
51 practically perfect association between the Trio-OptiTrack™ system and the different devices
52
53
54 was observed (r range = 0.947-0.995; P < 0.001) with the only exception of the Beast™
55
56 sensor (r = 0.765; P < 0.001) (Figure 3).
57
58
59
60
61
62
63
64
12
65
Feasibility of seven devices to measure velocity
the present study (T-Force™) has been frequently used to validate other devices that were
1
2 designed to measure movement velocity (2,13,14). The preferential use of linear velocity
3
4
5 transducers as compared to linear position transducers could be justified because the direct
6
7 measurement of movement velocity is expected to provide more accurate velocity outputs
8
9
10 (30). However, there is scarce information regarding the comparison of the reliability of
11
12 velocity outputs between linear velocity and linear position transducers. Contrary to our
13
14
15 hypothesis, the linear velocity transducer (i.e., T-Force™) was not more reliable than the two
16
17 linear position transducers used in the present study (i.e., Chronojump™ and Speed4Lift™).
18
19
In addition, to the best our knowledge, this is the first scientific article that has provided data
20
21
22 regarding the validity of the Chronojump™ and Speed4Lift™. It should be noted that the
23
24 Speed4Lift™ was the most reliable device and the only device that did not report
25
26
27 heteroscedasticity of errors with respect to the Trio-OptiTrack ™ system. These results
28
29 together with its lower price ( 340 US dollars) and excellent portability (the software is
30
31
32 installed in a smartphone app that is wireless connected with the hardware) place the
33
34 Speed4Lift as an accurate, cost-effective and practical device for the measurement of
35
36
37 movement velocity. It is worth noting that the heteroscedasticity of errors observed for the
38
39 other devices compromises their interchangeability. The T-force™ and PowerLift™ showed
40
41
42
progressively larger values than the Trio-OptiTrack™ with increasing velocities, while
43
44 opposite results were observed for the Chronojump™, Velowin™, PUSH™ band, and Beast
45
46 Sensor™.
47
48
49
50
51 The velowin™ has been recently developed as a more affordable and practical device
52
53
54 to measure movement velocity during resistance training exercises. However, to date, only
55
56 two studies have examined its reliability and concurrent validity to measure movement
57
58
59 velocity. García-Ramos et al. (11) found a comparable reliability (CV = 4.29-4.60%) and
60
61
62
63
64
14
65
Feasibility of seven devices to measure velocity
high validity (r = 0.97-0.98) between the T-Force™ and Velowin™ systems during the free-
1
2 weigh back squat exercise. Laza-Cagigas el al. (21) also observed a high validity (concordant
3
4
5 correlation coefficient = 0.96) and reliability (CV = 7.3% and ICC = 0.97) of the Velowin™
6
7 system to measure mean velocity during the free-weigh back squat exercise with respect to a
8
9
10 3D system. In line with these findings, our results showed a high and comparable reliability
11
12 and validity of the Velowin™ system as compared to the data provided by linear
13
14
15 position/velocity transducers during the bench press exercise performed in a Smith machine.
16
17 More recently, an affordable ( 11 US dollars) smartphone application named PowerLift™
18
19
20 has been made commercially available to measure movement velocity. The results of this
21
22 study corroborate the findings of Balsalobre-Fernández et al. who showed a comparable
23
24
25
reliability (ICC = 0.93-0.99) and high validity (r = 0.94-0.98) of PowerLift™ with respect to
26
27 the data provided by a linear position transducer (SmartCoach Europe, Stockholm, Sweden)
28
29 during free-weight exercises (3,4). Although it should be acknowledged that it is possible to
30
31
32 obtain accurate measurements of mean velocity with PowerLift™, the necessity of
33
34 individually select the start and end points of each repetition may be unpractical when many
35
36
37 athletes need to be assessed. In addition, the no provision of real-time velocity feedback
38
39 should also be considered as a limitation of compared to the other devices analysed in the
40
41
42
present study (39).
43
44
45
46 Wearable technology such as the PUSH™ band and Beast™ sensor are being
47
48
49 increasingly used in the strength and conditioning field. These two IMUs have been recently
50
51 validated to measure movement velocity during a variety of resistance training exercises
52
53
54 (2,3,5,36). In line with our results, the PUSH™ band provided highly valid measurements of
55
56 mean velocity when the data of several loads were combined for the analysis during the
57
58
59 bench press performed in a Smith machine as well as during the free-weight shoulder press
60
61
62
63
64
15
65
Feasibility of seven devices to measure velocity
and biceps arm curl exercises (r > 0.86) (2,36). However, the validity of the PUSH band
1
2 seems to be compromised when the data of individual loads are analysed separately (specially
3
4
5 for loads 80% of 1RM) (5). The PUSH™ band device should be therefore considered with
6
7 some caution given the controversial results and the lower reliability reported in the present
8
9
10 study. Regarding the Beast™ sensor, to the best of our knowledge, only one study has
11
12 investigated the reliability and validity of this device with respect to the data provided by a
13
14
15 linear position transducer (SmartCoach Europe, Stockholm, Sweden), reporting both a very
16
17 high reliability (ICC > 0.95) and validity (r > 0.98) during the bench press, full-squat and
18
19
20 hip-thrust exercises (3). On the contrary, our results suggest that the Beast™ sensor is the
21
22 least reliable and valid device among all the commercially available devices analysed in the
23
24
25
present study. It is plausible that the lower sampling frequencies of the IMUs or the need to
26
27 integrate acceleration-time data to obtain velocity values could have caused their lower
28
29 reliability (5,28). Therefore, more evidence about the feasibility of wearable technology in
30
31
32 the field of velocity-based resistance training is needed.
33
34
35
36
37 Several limitations and directions for future research should be considered. First, it
38
39 should be noted that the mean velocity is not the only velocity variable used in practice. The
40
41
42
mean propulsive velocity (i.e., average velocity from the start of the concentric phase until
43
44 the acceleration of the bar is lower than gravity) and maximum velocity (i.e., maximum
45
46 instantaneous velocity value reached during the concentric phase) have been commonly
47
48
49 recommended for training and testing (8,35). Therefore, future studies should consider
50
51 expanding the analysis of the reliability and validity of different devices to these variables. It
52
53
54 should be noted that the mean velocity was the only variable analysed in the present study
55
56 since it was the only common variable for the eight devices analysed. Second, the PUSH™
57
58
59 band was placed on the subject’s forearm, while the other devices were attached to the
60
61
62
63
64
16
65
Feasibility of seven devices to measure velocity
reliability and validity and, consequently, more caution should be taken when using PUSH™
1
2 band and Beast™ sensor devices for implementing velocity-based resistance training
3
4
5 programs.
6
7
8
9
10 REFERENCES
11
12 1. Atkinson, G and Nevill, AM. Statistical methods for assessing measurement error
13
14
15 (reliability) in variables relevant to sports medicine. Sport Med 26: 217–238, 1998.
16
17 2. Balsalobre-Fernández, C, Kuzdub, M, Poveda-Ortiz, P, and Campo-Vecino, JD. Validity
18
19
and reliability of the push wearable device to measure movement velocity during the
20
21
22 back squat exercise. J Strength Cond Res 30: 1968–1974, 2016.
23
24 3. Balsalobre-Fernandez, C, Marchante, D, Baz-Valle, E, Alonso-Molero, I, Jimenez, SL, and
25
26
27 Munoz-Lopez, M. Analysis of wearable and smartphone-based technologies for the
28
29 measurement of barbell velocity in different resistance training exercises. Front
30
31
32 Physiol 8: 649, 2017.
33
34 4. Balsalobre-Fernández, C, Marchante, D, Muñoz-López, M, and Jiménez, SL. Validity and
35
36
37
reliability of a novel iPhone app for the measurement of barbell velocity and 1RM on
38
39 the bench-press exercise. J Sports Sci 36: 64–70, 2018.
40
41 5. Banyard, HG, Nosaka, K, Sato, K, and Haff, GG. Validity of various methods for
42
43
44 determining velocity, force and power in the back squat. Int J Sports Physiol Perform
45
46 12: 1170–1176, 2017.
47
48
49 6. Conceicao, F, Fernandes, J, Lewis, M, Gonzalez-Badillo, JJ, and Jimenez-Reyes, P.
50
51 Movement velocity as a measure of exercise intensity in three lower limb exercises. J
52
53
54 Sports Sci 34: 1099–1106, 2016.
55
56 7. Cormack, SJ, Newton, RU, McGuigan, MR, and Doyle, TLA. Reliability of measures
57
58
obtained during single and repeated countermovement jumps. Int J Sports Physiol
59
60
61
62
63
64
18
65
Feasibility of seven devices to measure velocity
772–781, 2014.
1
2 16. González-Badillo, JJ and Sánchez-Medina, L. Movement velocity as a measure of
3
4
5 loading intensity in resistance training. Int J Sports Med 31: 347–352, 2010.
6
7 17. Harris, NK, Cronin, J, Taylor, KL, Boris, J, and Sheppard, J. Understanding Position
8
9
10 Transducer Technology for Strength and Conditioning Practitioners. Strength Cond J
11
12 32: 66–79, 2010.
13
14
15 18. Hopkins, WG, Marshall, SW, Batterham, AM, and Hanin, J. Progressive statistics for
16
17 studies in sports medicine and exercise science. Med Sci Sports Exerc 41: 3–13, 2009.
18
19
19. Hopkins, W. Calculations for reliability (Excel spreedsheet). A new view of statistics,
20
21
22 2000. Available from: http://www.sportsci.org/resource/stats/relycalc.html#excel.
23
24 21. Bryan, JM, Patrick A., I, Stephen P., S, Mann, J, Ivey, P, and Sayers, S. Velocity-Based
25
26
27 Training in Football. Strength Cond J 37: 52–57, 2015.
28
29 21. Laza-Cagigas, R, Goss-Sampson, M, Larumbe-Zabala, E, Termkolli, L, and Naclerio, F.
30
31
32 Validity and reliability of a novel optoelectronic device to measure movement
33
34 velocity, force and power during the back squat exercise. J Sports Sci, 2018. Epub
35
36
37
ahead of print.
38
39 22. Lorenzetti, S, Lamparter, T, and Luthy, F. Validity and reliability of simple measurement
40
41 device to assess the velocity of the barbell during squats. BMC Res Notes 10: 707,
42
43
44 2017.
45
46 23. Loturco, I, Kobal, R, Moraes, JE, Kitamura, K, Cal Abad, CC, Pereira, LA, and
47
48
49 Nakamura, FY. Predicting the maximum dynamic strength in bench press: the high
50
51 precision of the bar velocity approach. J Strength Cond Res 31: 1127–1131, 2017.
52
53
54 24. Loturco, I, Pereira, LA, Winckler, C, Santos, WL, Kobal, R, and McGuigan, M. Load-
55
56 velocity relationship in national paralympic powerlifters: a case study. Int J Sports
57
58
Physiol Perform, 2018. Epub ahead of print.
59
60
61
62
63
64
20
65
Feasibility of seven devices to measure velocity
25. McMaster, DT, Gill, N, Cronin, J, and McGuigan, M. A brief review of strength and
1
2 ballistic assessment methodologies in sport. Sports Med 44: 603–623, 2014.
3
4
5 26. Munoz-Lopez, M, Marchante, D, Cano-Ruiz, MA, Chicharro, JL, and Balsalobre-
6
7 Fernandez, C. Load-, force-, and power-velocity relationships in the prone pull-up
8
9
10 exercise. Int J Sports Physiol Perform 12: 1249–1255, 2017.
11
12 27. Nagymáté, G and Kiss, RM. Application of OptiTrack motion capture systems in human
13
14
15 movement analysis. A systematic literature review. Recent Innov Mechatronics 5,
16
17 2018.
18
19
28. Orange, ST, Metcalfe, JW, Liefeith, A, Marshall, P, Madden, LA, Fewster, CR, and
20
21
22 Vince, RV. Validity and reliability of a wearable inertial sensor to measure velocity
23
24 and power in the back squat and bench press. J Strength Cond Res, 2018. Epub ahead
25
26
27 of print.
28
29 29. Pareja-Blanco, F, Rodríguez-Rosell, D, Sánchez-Medina, L, Sanchis-Moysi, J, Dorado,
30
31
32 C, Mora-Custodio, R, et al. Effects of velocity loss during resistance training on
33
34 athletic performance, strength gains and muscle adaptations. Scand J Med Sci Sports
35
36
37
27: 724–735, 2017.
38
39 30. Pérez-Castilla, A, Feriche, B, Jaric, S, Padial, P, and García-Ramos, A. Validity of a
40
41 linear velocity transducer for testing maximum vertical jumps. J Appl Biomech 33:
42
43
44 388–392, 2017.
45
46 31. Pérez-Castilla, A, García-Ramos, A, Padial, P, Morales-Artacho, A, and Feriche, B.
47
48
49 Load-velocity relationship in variations of the half-squat exercise: Influence of
50
51 execution technique. J Strength Cond Res, 2017. Apub ahead of print.
52
53
54 32. Sakamoto, A, Sinclair, PJ, and Naito, H. Strategies for maximizing power and strength
55
56 gains in isoinertial resistance training: Implications for competitive athletes. J Sports
57
58
Med Phys Fitness 5: 153–166, 2016.
59
60
61
62
63
64
21
65
Feasibility of seven devices to measure velocity
FIGURE LEGEND
1
2 Figure 1. Distribution of the measurement devices during the testing protocol. (1) Trio-
3
4
5 OptiTrack™, (2) T-Force™, (3) Chronojump™, (4) Speed4Lift™, (5) Velowin™, (6)
6
7 PowerLift™, (7) PUSH™ band, and (8) Beast™ sensor.
8
9
10
11
12 Figure 2. Bland-Altman plots for the measurement of mean velocity between the Trio-
13
14
15 OptiTrack™ system and the seven commercially available devices: T-Force™ (upper-left
16
17 panel), Chronojump™ (upper-right panel), Speed4Lift™ (upper middle-left panel),
18
19
Velowin™ (upper middle-right panel), PowerLift™ (lower middle-right panel), PUSH™
20
21
22 band (lower middle-left panel), and Beast™ sensor (lower panel). Each plot depicts the
23
24 averaged difference and 95% limits of agreement (dashed lines), along with the regression
25
26
27 line (solid line).
28
29
30
31
32 Figure 3. Relationship of mean velocity values between the Trio-OptiTrack™ system and the
33
34 seven commercially available devices: T-Force™ (upper-left panel), Chronojump™ (upper-
35
36
37
right panel), Speed4Lift™ (upper middle-left panel), Velowin™ (upper middle-right panel),
38
39 PowerLift™ (lower middle-right panel), PUSH™ band (lower middle-left panel), and
40
41 Beast™ sensor (lower panel). The Pearson’s correlation coefficient (r) was calculated using
42
43
44 the log-transformation since the assumption of homoscedasticity was violated.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
23
65
Table
Table 1. Reliability of mean velocity values obtained from the Trio-OptiTrack™ method and
seven commercially available devices at different loads during the bench press exercise.
CV, coefficient of variation; ICC, intraclass correlation coefficient; 95% CI, 95% confidence
interval.
Figure 1_R1
(5)
(7)
(1)
(8)
(2)
0.15 0.15
Systematic bias ± random error: -0.01 ± 0.03 ms-1 Systematic bias ± random error: -0.03 ± 0.03 ms-1
r² = 0.266 r² = 0.508
0.10 0.10
0.05 0.05
0.00 0.00
-0.05 -0.05
-0.10 -0.10
-0.15 -0.15
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Mean Trio-OptiTrack™ and T-force™ Mean Trio-OptiTrack™ and Chronojump™
Differences (Trio-OptiTrack™ – Speed4Lift™)
0.05 0.05
0.00 0.00
-0.05 -0.05
-0.10 -0.10
-0.15 -0.15
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Mean Trio-OptiTrack™ and Speed4Lift™ Mean Trio-OptiTrack™ and Velowin™
Differences (Trio-OptiTrack™ – PUSH band™)
Differences (Trio-OptiTrack™ – PowerLift™)
0.15 0.30
Systematic bias ± random error: -0.01 ± 0.05 ms-1
r² = 0.437
0.10 0.20
0.05 0.10
0.00 0.00
-0.05 -0.10
-0.15 -0.30
0.20 0.40 0.60 0.80 1.00 0.15 0.35 0.55 0.75 0.95
Mean Trio-OptiTrack™ and PowerLift™ Mean Trio-OptiTrack™ and PUSH™ band
Differences (Trio-OptiTrack™ – Beast sensor™)
0.75
0.50
0.25
0.00
-0.25
-0.50
Systematic bias ± random error: 0.05 ± 0.21 ms-1
r² = 0.426
-0.75
0.05 0.35 0.65 0.95 1.25
Mean Trio-OptiTrack™ and Beast™ sensor
Figure 3_R1
1.00 1.00
y = 1.069x - 0.051 y = 0.881x + 0.046
r = 0.995; P < 0.001 r = 0.991; P < 0.001
Trio-OptiTrack™ (ms-1)
Trio-OptiTrack™ (ms-1)
0.80 0.80
0.60 0.60
0.40 0.40
0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
T-force™ (ms-1) Chronojump™ (ms-1)
1.00 1.00
0.60 0.60
0.40 0.40
0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Speed4Lift™ (ms-1) Velowin™ (ms-1)
1.00 1.00
y = 1.205x – 0.130 y = 0.822x + 0.187
r = 0.970; P < 0.001 r = 0.947; P < 0.001
Trio-OptiTrack™ (ms-1)
Trio-OptiTrack™ (ms-1)
0.80 0.80
0.60 0.60
0.40 0.40
0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.00 0.20 0.40 0.60 0.80 1.00
PowerLift™ (ms-1) PUSH™ band (ms-1)
1.00
0.80
Trio-OptiTrack™ (ms-1)
0.60
0.40
y = 0.456x + 0.341
r = 0.765; P < 0.001
0.20
0.00
0.00 0.30 0.60 0.90 1.20 1.50
Beast™ sensor (ms-1)