Participants. Ten healthy (free from lower limb injury for past 6 months) well trained male athletes volunteered to take part in this study (age 28 ± 5yrs, weight 73 ± 8 kg, height 180 ± 6 cm). Informed consent was obtained for each participant, following approval of the local ethics committee (Anti-Doping Lab Qatar Approval number #E2018000272). Participants regularly ran at speeds greater than 24 km/hr and in the 3-months leading into the study, ran on average 31 ± 15 km per week (self-reported). All participants were familiar with treadmill running usually completing a minimum of one treadmill session per week in the 3months prior.
Equipment. Plantar loading parameters were measured using an in-shoe load monitoring device (Loadsol®, Novel, Munich, Germany). Each Loadsol® insole consists of two capacitive force sensors that transmit data over Bluetooth to a smartphone or tablet. Force sensor insoles were placed inside the participants own preferred running shoes in the appropriate size. No participants used orthotic supports. Insole calibration followed manufacturers guidelines (Novel, Munich, Germany) with calibration accepted if a bodyweight ± 5% of the athletes’ bodyweight was achieved at single leg stance with the insoles fully loaded. Insole resolution was set at 5/10Newtons for a range of 0-2550N and a sample rate of 200 Hz using the Loadsol® App (version 1.5.10) on an Apple Ipad Pro 9.7 inch (Apple, Cupertino, USA). Here, we use the most recent generation of Loadsol® insoles with a sample rate of 200 Hz which demonstrated improved validity (ICC 0.76–0.98) over the previous generation of insoles (100 Hz) when compared with a force plate sampling at 1920 Hz [15]. Loadsol® insoles sampling at 200 Hz underestimate vertical force measurements in a reliable way when compared to force plates (95% limits of agreement 0.02 to 0.69BW) [15, 12].
Warm-up protocol. Participants were fitted with the correct sized AlterG® shorts. A calibration protocol according to the manufacturer instructions was followed whereby the athlete stands with arms folded across their chest while the bodyweight of the athlete is measured on the treadmill deck (G-trainer pro 2.0, AlterG®, California USA). All participants used the same warm-up protocol: Walk for 5 minutes at 5 km/hr at 100% bodyweight (No AlterG® assisted BW support), Run for 3 minutes at 10 km/hr at 100% of BW. Followed by 2 × 10 second efforts at 21 km/hr and 2 × 10 second efforts at 24 km/hr at 100% BW (with 30 seconds static recovery in between efforts), in order to familiarize themselves with getting on and off the treadmill at high speed.
Testing Protocol Following warm-up participants ran at varying combinations of AlterG® indicated BW support (60%, 80%, and 100% BW), speed (12 km/hr, 15 km/hr, 18 km/hr, 21 km/hr, and 24 km/hr), and gradients (-15% decline, -10, -5, 0, + 5, +10 + 15% incline) for approximately 60 seconds per trial. Sum of total running trials was 78 with all possible combinations. Each of these combinations of speed, gradient and BW was block randomised a priori using online software (www.randomizer.org). A recovery period was set at a minimum of 45 seconds between each trial. All data was collected at a single visit for each participant.
For the downhill running trials, participants faced ‘backwards’ in the AlterG® treadmill and the belt was run in reverse. Hence, the inline function with belt in reverse direction can be used as a decline when facing away from the usual running direction. Top speed for the treadmill (G-trainer pro 2.0, AlterG®, California USA) in reverse was 15 km/hr, therefore, all decline conditions were conducted at two running speeds of 12 &15 km/hr.
For each trial condition participants were instructed to run until they felt comfortable, and then indicate the point where their gait felt “normal”. Loadsol® insole data was then collected at 200 Hz for a minimum of 6 stance phase foot contacts of both feet. Threshold of 40N was set to identify when stance phase commenced to decrease any signal noise associated with treadmill running.
Statistical & data Analysis.
All data were processed using custom scripts( https://github.com/JuliusWelzel/AlterG-loadsol) for MATLAB (Version 9.6; MathWorks, Natick, MA, USA). Maximum plantar force (Fmax) for each foot were extracted respectively from the time of stance and averaged for subsequent analysis over a minimum of six footfalls for each foot (twelve total). Maximum force was normalised to participants bodyweights to aid comparison across the group. Outliers in the data excluded elements more than 1.5 interquartile ranges above the upper quartile or below the lower quartile [17]. Multiple linear regression was used to reveal the relationship between running speed, percentage body weight and normalized maximal plantar force as outcome variable. To understand the effect of different gradients during running on the loading forces, another multiple linear regression was conducted with running speed, AlterG® assisted BW support, and gradient as regressors. Post-hoc analysis used repeated measures ANOVA with Bonferroni correction and subsequent pairwise comparisons and effect size calculations. Level of significance was set a priori at p = 0.05. Effect sizes (cohen’s d) were reported as small, medium, large, and very large when they reached 0.2, 0.5, 0.8, 1.2 respectively [18, 19]. The maximum plantar force data collected by the Loadsol® insoles are reported in units of BW (times bodyweight). The indicated AlterG® bodyweight support on the treadmill is reported as percentage of bodyweight (%BW). Treadmill incline or decline is reported as a gradient (%). Contact time is reported in milliseconds (ms).