Concurrent validity of the portable gFlight system compared to a force plate to measure jump performance variables

Objective. Lower-limb strength and power is commonly assessed indirectly by measuring jump performance. A novel portable system (gFlight) that can be used in applied settings provides measures of jump performance. The aim of this study was to validate jump performance measures provided by the gFlight to those provided by a force plate. Approach. Thirty-six participants performed three countermovement jump (CMJ) and drop jump (DJ) trials. Jump height (JH), contact time, and reactive strength index (RSI) were simultaneously recorded by a force plate and gFlight sensors to assess concurrent validity. Main results. The gFlight provided significantly higher measures of JH during the CMJ (mean: +8.79 ± 4.16 cm, 95% CI: +7.68 to 9.90 cm, P < 0.001) and DJ (mean: +4.68 ± 3.57 cm, 95% CI: +3.73 to 5.63 cm, P < 0.001) compared to the force plate. The gFlight sensors displayed significantly higher measures of RSI (mean: +0.48 ± 0.39 m·s−1, 95% CI: +0.37 to 0.58 m·s−1, P < 0.001) and lower measures of contact time (mean: −0.036 ± 0.028 s, 95% CI: −0.044 to −0.029 s, P < 0.001) during the DJ compared to the force plate. The bias displayed by the gFlight for JH, contact time and RSI measures are reduced using corrective equations. Significance. The gFlight sensors are a cost-effective, portable measurement system with high concurrent and ecological validity for the objective measurement of jump performance in applied settings. Corrective equations should be used to reduce measurement biases so comparisons can be made to force plate measurements of jump performance.


Introduction
Lower-limb power is commonly assessed indirectly by measuring jump height (JH) performance during vertical jumping tasks such as the countermovement jump (CMJ) and drop jump (DJ) (Bosco et al 1983, Harman et  Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Markovic et al 2004). The measurement of JH is a frequently used method to assess and monitor physical performance and adaptations by coaches and researchers (Coutts et al 2007, Twist andHighton 2013), along with being one of the most prevalent activities performed in a wide range of sports (Glatthorn et al 2011). Assessing lowerlimb performance during jumping tasks provides coaches and researchers with information relating to the utilisation of the stretch-shortening cycle (SSC) and reactive strength index (RSI) during the CMJ and DJ, respectively (Flanagan andComyns 2008, Twist andHighton 2013).
Force plates are considered the 'gold standard' to measure jump performance (Cronin et al 2004, Buckthorpe et al 2012. Force plates are mechanical systems that provide measurements of ground reaction forces and moments involved with human movement (Lamkin-Kennard and Popovic 2019); however, these are often expensive (∼£20k), bulky and require specialist software to collect and analyse data. The use of force plates to measure JH where access to laboratory facilities are limited are therefore impractical, however, applied practitioners still need to assess and monitor the physical performance and readiness of the athletes they support.
In order to make traditional lab-based performance tests more accessible, advances in technology have provided applied practitioners and athletes with access to field-based measures of JH that can be used in their own environments. These include contact mats (Just Jump system), velocity systems (GymAware), linear position transducers (MyoTest, Vertec), optical photoelectric cells (OptoJump), and mobile phone applications (MyJump) (Cronin et al 2004, Glatthorn et al 2011, Balsalobre-Fernández et al 2015, Hojka et al 2018. These field-based alternatives, however, all use different software and calculations to provide JH measurements meaning results can vary depending on the system used. With portable and wearable technologies increasing in popularity, more research is being published to evaluate the reliability and validity of these measurement systems (Bubanj et al 2010, Yingling et al 2018. Recently, in 2018, a novel portable measurement system was developed (Exsurgo gFlight v2,) that can fit into a small bag and connects to a free downloadable smartphone application (gTechAMS, Exsurgo Technologies, LLC) via Bluetooth. The system consists of two photoelectric boxes; one transmitter and one receiver that are placed at a maximum of 5.8 m apart at floor level. The gFlight measures JH via time in air, CT and RSI, with participants instructed to stand with their fifth metatarsal in line with the beam. The portability and relatively low price ($399) of the gFlight makes this system an accessible option for applied practitioners, as well as improving the ecological validity of the measurements taken. The validity of the gFlight, however, is unknown, with no studies currently published evaluating the validity of the measures provided by the gFlight system against those provided by the 'gold standard' force plate.
The aim of this study is to provide a novel evaluation of the concurrent validity of the gFlight compared to the 'gold standard' force plate to measure JH, CT, and RSI during a CMJ and DJ. The evaluation of this novel measurement system will provide researchers and practitioners with knowledge as to the validity of the measures provided by the gFlight for the first time.

Method Participants
With institutional ethics approved by Northumbria University Research and Ethics Committee, 36 young healthy adults (27 male, nine female) participated. The age, stature and mass of participants, reported as mean±SD, were 22.0±4.4 yrs; height: 1.75±0.08 m; 74.87±11.88 kg. The inclusion criteria for participation in this study were that participants had to be aged 18-35 years, and free from physical limitations or musculoskeletal injuries that could affect their ability to perform the testing procedures. Participants were excluded if they had an injury to the lower limb or had any condition that would affect jumping performance. Participants represented a wide range of abilities and training status from recreational to highly trained, participating in 1.5 to 14 h of moderate to strenuous physical activity per week, as defined in the American College of Sports Medicine (ACSM) Physical Activity Guidelines (Powell et al 2018). This was to ensure the gFlight system could be validated across a wide range of JHs. All participants were asked to refrain from strenuous exercise in the 24 h prior to testing. Testing procedures were conducted on two separate occasions separated by 1-week at the same time of day (1300-1700), with participants wearing the same pair of their own athletic shoes with cushioning for all trials.

Study design
All participants performed three maximal trials of the CMJ and the DJ with hands placed on the hips throughout, following a standardized 10 min warm-up. Data for each trial were simultaneously recorded using a floor integrated force plate (AMTI Biovac 1100, Watertown, MA, USA) (criterion instrument) and a pair of Exsurgo gFlight sensors (Exsurgo, Virginia, USA) (practical instrument) to assess the concurrent validity of this latter system, with the averages of the three CMJ and DJ trials used for further analysis. The dependent variables were JH of the CMJ and DJ, and the contact time and RSI of the DJ. The independent variables were the measurement tools; specifically, the force plate as the gold standard criterion measure, and the gFlight sensors as the practical experimental measure.

Procedures
Upon arrival to the laboratory, a full explanation of the experimental protocol and procedures was provided to participants. Following this, participants completed a standardized 10 min warm-up led by the principal investigator following the raise, activate, mobilise and potentiate protocol (Jeffreys 2007) consisting of movements similar to those detailed in similar previous studies (Glatthorn et al 2011, Attia et al 2017, Driller et al 2017. At the end of the warm-up participants performed three submaximal CMJ and three submaximal DJ at 50, 75, and 90% perceived maximum effort, familiarizing participants with the jump protocols. Each participant subsequently performed the three maximal CMJ trials and the three maximal DJ trials on the force plate and between the gFlight sensors. All jump trials were separated by 30 s of rest, with 2 min between the types of jumps. For the standardisation of all jumps, participants kept their hands on their hips throughout the entire movement and were instructed to jump vertically with as little horizontal displacement as possible and land in the same place as take-off. For the CMJ, participants stood in an upright position with feet approximately shoulder width apart. From this position, participants were instructed to squat to approximately 90°of knee flexion as fast as possible before then jumping as high as possible. For the DJ, participants stood in an upright position with feet shoulder width apart on a 0.30 m box before stepping forwards off of the box. Upon contact with the ground, participants jumped as high as possible, as quickly as possible, attempting to achieve the greatest JH with the least ground contact time (Young et al 1995). Jump trials not meeting these procedures were deemed invalid and participants repeated the trial.
The Exsurgo gFlight sensors were placed at the extremities of the force platform without touching it, in a parallel and horizontal position to one another at a distance of 0.56 m (figure 1). The Exsurgo gFlight sensors were connected via Bluetooth to an iPhone SE (Apple Inc., USA) to record jump trials on the Exsurgo gtech application, with all dependent variables (JH , contact time, and RSI) automatically calculated. The force plate (AMTI Biovac 1100, Watertown, MA, USA) was integrated into the floor to measure the vertical ground reaction force (VGRF) during jumping at a sampling rate of 2000 Hz (figure 1). The force time trace recorded for each trial was used to directly calculate all dependent variables (JH , contact time, and RSI). Contact times and flight times were obtained using a threshold of >10 N to determine contact and <10 N to determine flight (Healy et al 2016). JH from force plate data was estimated as 9.81 × flight time 2 /8 (Bosco et al 1983). The RSI was calculated by dividing the JH by the contact time (Flanagan andComyns 2008, Markwick et al 2015).

Statistical analyses
All data are presented as mean±standard deviation (SD). Normality was assessed by visual inspection of box plots for all dependent variables before analyses. The dependent variables obtained from the three CMJ trials and the three DJ trials performed by each participant were averaged for further analyses. Paired samples t-tests (and associated 95% confidence intervals) were used to detect systematic differences (also referred to as bias) between tools (validity) for all dependent variables. Concurrent validity between measurement tools for all dependent variables were examined using bivariate linear regression, CV, and Pearson's correlation coefficient (r). The SEE was calculated to assess the accuracy of the predictive equation from the linear regression. The coefficient of determination (R 2 ) was calculated to demonstrate the relationship between the dependent variables measured from the two measurement tools. Effect sizes (d) were calculated to determine the magnitude of differences between the two measurement tools for all dependent variables. A modified scale was used for the interpretation of d; d<0.2 as trivial, 0.2-0.6 as small, 0.6-1.2 as moderate, 1.2-2.0 as large, 2.0-4.0 as very large, and >4.0 as extremely large (Hopkins 2010). The magnitude of correlation (r) was interpreted as; <0.10 as trivial, 0.10-0.30 as small, 0.30-0.50 as moderate, 0.50-0.70 as large, 0.70-0.90 as very large, and 0.90-1.00 as almost perfect (Hopkins et al 2009). All analyses were performed using the Microsoft Excel (2013) statistical package using a spreadsheet for validity (Hopkins 2015). Statistical significance was accepted when P<0.05.

Jump height
The gFlight sensors demonstrated a very large agreement with the force plate for the measurement of JH in both the CMJ (r=0.83) and the DJ (r=0.83). Despite this agreement, the gFlight displayed a significant systematic bias with higher measures of JH provided in comparison to the force plate during the CMJ (mean: +8.79±4.16 cm, 95% CI: +7.68 to 9.90 cm, d: 1.25, P<0.001) and the DJ (mean: +4.68±3.57 cm, 95% CI: +3.73 to 5.63 cm, d: 0.83, P<0.001) (table 1). The systematic bias demonstrated between the two measurement tools increased with increasing JH, as predicted by the linear regression equations for both the CMJ (figure 2(A)) and the DJ (figure 3(A)); with 69% and 68% of the variance in JH explained by the respective equations. The SEE was ±3.80 cm during the CMJ and ±2.81 cm during the DJ. The CV describing the concurrent validity between measurement tools were 13.60% for the CMJ and 13.40% for the DJ (table 1).  Correcting the gFlight measurement of JH using the linear regression equations for the CMJ: corrected CMJ height=0.7595×raw gFlight JH+0.6306; and the DJ: corrected DJ height=0.647×raw gFlight JH+4.7173; reduced the significant systematic bias displayed between the two measurement tools in both the CMJ (mean: 0.00±3.77 cm, 95% CI: −1.00 to 1.01 cm, d<0.001, P=0.99) and the DJ (mean: 0.00± 2.78 cm, 95% CI: −0.74 to 0.74 cm, d<0.001, P=0.99) (table 2). The corrected gFlight JH measures demonstrated very large agreement with the force plate in both the CMJ (r=0.83) and the DJ (r=0.83), with the linear regression equations displaying a nearly perfect relationship in the CMJ (y=1x+6×10 -6 ; figure 2(B)) and the DJ (y=1x -0.0001, figure 3(B)).

Contact time and RSI
The gFlight sensors displayed a significant systematic bias for the measurement of contact time and RSI, with a lower measure of contact time (mean: −0.036±0.028 s, 95% CI: −0.044 to −0.029 s, d: −0.75, P<0.001) and a higher measure of RSI (mean: +0.48±0.39 m·s −1 , 95% CI: +0.37 to 0.58 m·s −1 , d: 0.97, P<0.001) provided compared to the force plate (table 1). Pearson correlation values demonstrated very large agreement between measurement tools for both contact time (r=0.83) and RSI (r=0.75). The systematic bias displayed by the gFlight sensors compared to the force plate for the measurement of contact time was consistent as predicted by the linear regression equation, with a SEE of ±0.028 s and the equation explaining 69% of the variance observed ( figure 4(A)). The systematic bias observed between the two measurement tools increased with increasing RSI as predicted by the linear regression equation, with a SEE of ±0.25 m·s −1 and the equation explaining 56% of the variance observed ( figure 5(A)). The CV values describing the concurrent validity between measurement tools for contact time and RSI were 13.70% and 26.20%, respectively (table 1).

Discussion
The aim of this study was to evaluate the concurrent validity of the gFlight sensors in comparison to a force plate to measure JH, CT and RSI during a CMJ and DJ. This is the first study to evaluate the novel gFlight system to a 'gold standard' criterion force plate, providing practical information pertaining to the validity of the gFlight sensors for use in applied settings. The major findings from this study were that the gFlight system demonstrated strong concurrent validity compared to the force plate for all measures during the CMJ and DJ. Despite this, a significant systematic bias was displayed between the two measurement tools, as the gFlight provided higher measures of JH and RSI during the CMJ and DJ, respectively, with the observed bias increasing with increasing JH and RSI. Similarly, measurements of CT provided by the gFlight were systematically lower than those provided by the force plate, however the bias observed was consistent irrespective of the contact time measurement. Nevertheless, the gFlight demonstrated very large agreement for all measures (r values ranging between 0.75 and 0.83) between the gFlight and force plate. The use of corrective equations derived from the linear regression equations reduced the systematic bias observed between measurement tools for all measures, thereby making this a potentially valid measurement tool to use within applied settings. The higher systematic bias observed between the gFlight and force plate for the measurement of JH contrasts previous research evaluating the validity of similar systems using photoelectric cells (Optojump) to estimate JH, from the measurement of flight time. Differences between measures of JH using the Optojump are consistently reported to be systematically lower than force plate measures of JH, typically attributed to the photoelectric cells being raised off of the ground leading to lower measures of flight time and in turn JH (Glatthorn et al 2011, Castagna et al 2013, Attia et al 2017. The measurement of flight time from photoelectric cell devices is dependent upon the detection of take-off and landing (Glatthorn et al 2011, Castagna et al 2013. The detection area of the gFlight system is relatively small in comparison to the Optojump, therefore any horizontal displacement exhibited during the flight phase of a jump might affect the measurement of flight time due to the landing location being different to the take-off location (Attia et al 2017). The smaller detection area of the gFlight might therefore overestimate flight time due to differences in the detection of take-off and landing, and in turn the JH measure. In comparison, the Optojump system has a larger detection area, therefore any horizontal displacement exhibited during the flight phase of a jump will not affect the JH measure provided. This difference in the size of the detection area perhaps explains the contrasting biases observed compared to the force plate for the measurement of JH. Another field-based alternative to measure JH via flight time is a smartphone application, that reportedly provides a measure of JH similar to that provided by a force plate (mean bias=0.9±0.2 cm) (Driller et al 2017). Although the reported bias is lower than that shown here for the gFlight, the smartphone application relies on the user filming the jump trial at a suitable frame rate along with correctly identifying the take-off and landing frames for the calculation of flight time and hence JH (Driller et al 2017). The additional input required when using the smartphone application in comparison to the gFlight might reduce the systematic bias observed, however the gFlight offers a method to measure JH instantly without additional input, along with the presented corrective equations reducing the bias. Similarly, another alternative to force plates is the use of an accelerometer to measure JH via flight time, with the reported mean bias (3.6±0.1 cm) also less than the gFlight (Castagna et al 2013). The use of the accelerometer however requires specific and consistent placement on the participant for reliable JH measurements, along with specialist software to analyse the data. Furthermore, despite the accelerometer being a more cost-effective option than force plates, the price is still relatively higher than the gFlight system (Nuzzo et al 2011). When compared to other field-based alternatives for the measurement of JH, the gFlight demonstrates a higher systematic bias for the measurement of JH during both CMJ and DJ modalities (Nuzzo et al 2011, Buckthorpe et al 2012, Castagna et al 2013, Driller et al 2017. Nevertheless, the portability, low cost and accessibility might appeal to applied practitioners and researchers despite the greater systematic bias demonstrated compared to other field-based alternatives. With this in mind, the use of corrective equations presented herein can improve the validity of the gFlight system. The present findings show the corrective equations for CMJ JH (corrected CMJ height=0.7595×raw gFlight JH+0.6306) and DJ JH (corrected DJ height=0.647×raw gFlight JH+4.7173) lead to the large (CMJ JH: +8.79±4.16 cm, d=1.25) and moderate (DJ JH: +4.68±3.57 cm, d=0.83) systematic biases to be reduced to trivial (CMJ JH: 0.00±3.77, d=<0.001; DJ JH: 0.00±2.78 cm, d=<0.001) biases, effectively reducing the difference demonstrated between the force plate and gFlight. The gFlight sensors can therefore be considered valid measures of JH in both the CMJ and DJ with the use of the proposed corrective equations, which have been derived from a population of varied athletic ability.
The current study sought to evaluate measures of contact time and RSI provided by the gFlight during a DJ, as this information is relevant to practitioners attempting to assess the reactive SSC abilities of the athletes they support (Bosco et al 1983, Flanagan and Comyns 2008, Twist and Highton 2013. The RSI provides a measure of an athletes' ability to develop maximal force in minimal time through the utilisation of the fast SSC, derived from the measurement of JH divided by the ground contact time (Flanagan and Comyns 2008). The SSC consists of an eccentric muscle contraction immediately followed by a concentric muscle contraction, with a shorter time between these phases facilitating a greater ability to generate force due the ability to utilise the SSC (Bosco et al 1983, Flanagan andComyns 2008). The gFlight sensors provided systematically lower and higher measures of CT and RSI, respectively compared to the force plate. As RSI is calculated from JH and contact time (Flanagan andComyns 2008, Healy et al 2016), the higher JH and lower CT measures provided by the gFlight result in the higher RSI demonstrated in comparison to the force plate. The validity of CT and RSI measures from field-based measurement tools during a DJ is limited, as previous research has focussed primarily on vertical jumping tasks such as the CMJ or squat jump (Markovic et al 2004, Bubanj et  The differences in contact time and RSI measures provided by the gFlight system in comparison to the force plate are not dissimilar from the differences demonstrated by the aforementioned field-based alternatives. When compared to the reported differences in CT and RSI demonstrated by the Optojump (due to this system also utilising photoelectric cells), the gFlight does provide higher RSI measures and lower CT measures. This is most likely due to the size of the detection area, as previously explained. Nevertheless, in comparison to other field-based alternatives, the gFlight sensors offer a portable, time efficient and cost-effective option for applied practitioners and researchers alike to obtain objective measures of DJ performance. To allow comparisons of contact time and RSI measures to be made between the gFlight and force plate, the corrective equations presented in this study (corrected DJ contact time=0.9497×raw gFlight contact time+0.0458; corrected DJ RSI=0.4781×raw gFlight RSI+0.2994) can be used to reduce the systematic bias observed between the measurement tools. These equations can therefore be used to provide valid measures of CT and RSI in applied settings that have been derived from the gFlight sensors.

Limitations
The measurements of JH , contact time, and RSI provided by the gFlight in this study can be considered acceptable and valid when compared to the differences demonstrated by other validated field-based alternatives. The evaluation of the gFlight sensors, however, does not come without its limitations. The high CV values reported (13.50%-26.20%) are considered to be unacceptable according to previous studies reporting CV values <10% to be acceptable for biomechanical variables (Hunter et al 2004, Cormack et al 2008. The high variability observed in this study is most likely attributed to the mixed athletic ability of the participants, as demonstrated by the large range of scores for JH (CMJ: 25.76 to 55.94 cm; DJ: 12.96 to 41.27 cm), contact time (0.097 to 0.293), and RSI (0.54 to 3.84 m·s −1 ) measured by the gFlight sensors. It is also acknowledged that horizontal displacement can vary between participants when performing jumps, which combined with the small detection area of the gFlight sensors could potentially contribute further to the observed measurement variability, however this was not measured. Furthermore, this variability might have been present during participants perceived maximum effort warm-up trials, however, these jumps were not measured which is a possible limitation. Whilst we acknowledge the CV values can be considered unacceptable, the mixed athletic ability of the sample population allows the concurrent validity of the gFlight sensors to be tested across a wide range of JHs. A further limitation lies in the familiarity of the participants to perform the jump protocols. Despite familiarisation and instruction, there might still be inherent learning effects, especially for the performance of the DJ protocol for participants that do not perform such activities regularly, therefore contributing to the large variation observed. In addition, it is worth mentioning the number of trials where incomplete data was provided by the gFlight when participants performed their jumps. Of the 324 trials performed, the gFlight provided incomplete data on 6 occasions (1.85%), however, this low rate had no significant impact upon the ability to complete the tests and the subsequent data analyses. It is suggested future research examining the validity of the gFlight sensors should focus on populations in which jumping activities are performed regularly, such as basketball, volleyball and netball. Such research would therefore be able to evaluate if the systematic bias and variation observed in a mixed population is evident in trained athletic populations, along with if the corrective equations presented herein are applicable to these populations.

Conclusion
This study evaluated the concurrent validity of the novel gFlight sensors to provide measures of JH, contact time, and RSI during a CMJ and DJ in comparison to those provided by a 'gold standard' force plate. The gFlight sensors provided valid measures of the dependent variables in both jump modalities; however, systematic biases were demonstrated. Corrective equations should be used to reduce these biases and allow valid comparisons to be made to force plate measures of JH, CT and RSI during CMJ and DJ tasks. The gFlight sensors can therefore be considered a cost-effective, portable measurement system with high concurrent and ecological validity for the objective measurement of jump performance in applied settings.