Participants
Based on previous studies [5, 11, 13], a coefficient greater than 0.3 in the multiple regression model was anticipated to predict lower-limb extensor moments. To achieve a significance level (α) and statistical power (1 − β) of 0.05 and 0.8, respectively, 27 participants were needed. Therefore, the present study enrolled 28 participants (14 female and 14 male participants, age 22.8 ± 1.3 years, height 167.8 ± 8.0 cm, and body mass 58.8 ± 10.1 kg). Participants were excluded from this study if they reported pain during double-leg squatting, had any history of musculoskeletal injury within the prior 6 months, or had surgery of the lower extremities or the trunk. Written informed consent was obtained from each participant before participation. This study was approved by the Institutional Review Board of Faculty of Health Sciences, Hokkaido University (approval number: 19-72).
Procedures and data collection
Participants warmed up using a stationary bicycle ergometer at a self-selected pace for five minutes. Then, retroreflective markers were placed on the iliac crest, anterior and posterior superior iliac spines (ASISs and PSISs, respectively), lateral thigh, medial and lateral femoral epicondyle, lateral shank, medial and lateral malleoli, second metatarsal head and base, fifth metatarsal head and heel. Following a static standing trial, participants performed three sets of five consecutive double-leg squats. They squatted down until their thighs were parallel to the floor and then stood upright [15]. If their heels came off the floor, they were instructed to squat as deeply as possible without their heels coming off the floor. No specific instructions were given regarding trunk flexion. The descent and ascent phases were set to 2 s each using a metronome. Participants were asked to place one foot on an individual force plate with their feet shoulder width apart and to cross their arms over their chests. Two to three minutes of rest was allowed between each trial.
A motion capture system (Cortex version 5.0.1, Motion Analysis Corporation, Santa Rosa, CA, USA) with seven high-speed digital cameras (Hawk cameras, Motion Analysis Corporation) and two synchronized force plates (Type 9286, Kistler AG, Winterthur, Switzerland) were used for data recording. The sampling rates were set to 200 Hz for the marker trajectory data and 1000 Hz for the force plate data.
Data processing and reduction
Kinematic and kinetic analyses were performed using Visual3D (version 6, C-Motion, Inc., Germantown, MD, USA). Marker trajectories and force plate data were low-pass filtered using a fourth-order, zero-lag Butterworth filter with a 12 Hz cutoff frequency [4, 5]. The trajectory gaps of ASIS markers during the squatting position were filled using iliac crest and PSIS markers [16]. Trunk flexion and lower-limb joint angles were calculated using a joint coordinate system with the Cardan sequence. The trunk flexion angle was determined relative to the laboratory coordinate system. The knee extensor moment was calculated using an inverse dynamics approach, and the segment inertial properties were set according to a previous report [17]. The present study examined the knee extensor moment contribution because the absolute knee extensor moment can also be affected by the squatting speed and interlimb difference in weight bearing [1, 13, 18]. To determine the knee extensor moment contribution, the knee extensor moment was normalized to the total support moment, which was the sum of the hip, knee and ankle extensor moments (% total support) [3]. The COP position of each foot was calculated for AP direction. The AP direction was adjusted by the vector from the heel marker to the 2nd metatarsal head marker. The AP-COP position was expressed as the percentage of the foot length (% foot length) from the heel marker (0%) to the second metatarsal head marker (100%).
The middle three of the five consecutive squats were analyzed [4]. The knee extensor moment, ankle dorsiflexion angle, trunk flexion angle and AP-COP position at peak knee flexion were used in the subsequent statistical analysis [19]. Interlimb asymmetry was assessed using the LSI, which was calculated as the percentage of the value of the dominant limb to that of the nondominant limb. The dominant leg was determined as the side preferable for kicking a ball. All variables were averaged across the three squats of the three trials.
Statistical analysis
All data are presented as the mean and standard deviation (SD). Univariate regression analysis was conducted to confirm the linear relationship between the knee extensor moment contribution and the AP-COP position and ankle dorsiflexion and trunk flexion angles. In addition, the linear relationships between the AP-COP position, ankle dorsiflexion angle and trunk flexion angle were also examined. Then, multivariate regression analysis was performed to determine the predictive ability of AP-COP position independent of ankle dorsiflexion and trunk flexion angles. A regression model of ankle dorsiflexion with trunk flexion was also tested. Analysis was performed regarding the dominant side, nondominant side and LSI. A paired t test was also conducted to confirm differences in variables of interest between the dominant and nondominant sides. The statistical significance level was set at P < 0.05. These statistical analyses were performed using JMP Pro software (version 15, SAS Institute Inc., Cary, NC, USA).