J. Hum. Ergol., 43: 79-95, 2014
THE INFLUENCE OF LATERALITY ON DIFFERENT PATTERNS OF ASYMMETRICAL FOOT PRESSURE AND MUSCLEACTIVATION DURING A GAIT CYCLE IN MANUAL PUSHING KADEK HERI SANJAYA1, 3*, SOOMIN LEE2, YOSHIHIRO SHIMOMURA1 AND TETSUO KATSUURA1 Humanomics Laboratory, Department of Design Science, Graduate School of Engineering, Chiba University, Japan 2 Center for Environment, Health and Field Sciences, Chiba University, Chiba, Japan. 3 Research Centre of Electrical Power and Mechatronics, Indonesian Institute of Sciences, Bandung, Indonesia *E-mail:
[email protected] 1
ABSTRACT This study investigated laterality of manual pushing during a gait cycle by measuring pushing force, muscular activation and foot pressure. Subjects were 17 healthy young adult males; (11 right-handed [RH], and 6 left-handed [LH]). They pushed a force plate while walking on a treadmill at 1.5, 3, and 4 km/h. Electromyogram (EMG) data were collected bilaterally from the tibialis anterior, soleus, lumbar erector spinae and triceps brachii. To measure foot pressure, ten pressure sensors were attached bilaterally on five points of the sole. Symmetry assessment was performed by comparing bilateral data and cross-correlation function (CCF). Gait cycle duration was found to be symmetrical in all conditions. LH subjects demonstrated asymmetry in calcaneus contact duration to control ankle flexion, whereas RH were symmetrical. Velocity affected tibialis anterior muscle time lag and soleus muscle CCF coefficients, mainly in LH. We found that triceps brachii muscle CCF coefficients in LH subjects were affected by increasing velocity. Results indicated that LH and RH did not mirror each other, since both had distinct characteristics. Furthermore these asymmetries were not strictly associated with the preferred side, indicating that generalisation of preferred side to whole-body coordination should be avoided, since we could not separate one side from the other. Key words: handedness; footedness; EMG; cross-correlation function; time lag; foot contact duration INTRODUCTION
Manual materials handling is very common in industrial work due to limited space, changing activities, and inability of supporting equipment to move autonomously. Manual materials handling often involves pushing, as well as activities such as load lifting, load carrying and pulling. In pushing, unlike lifting, the load is actually supported by the floor instead of the body and the force is exerted to move the object (Mittal et al., 1997). Previous studies on manual pushing have focused on industrial ergonomics to reduce the musculoskeletal injury risk by investigating handle height (Resnick and Chaffin, 1995; Jansen et al., 2002; Hoozemans et al., 2007), the upper extremities (Voorbij and Steenbekkers, 2001) and trunk muscles load (Hoffman et al., 2007). Pushing is also performed in various activities of daily life; for example, babies perform pushing when learning to walk (Trettien, 1900), and people with balance deficiencies (Palisano et al., 2007) and older adults use walking aids that involve pushing. Pushing is also common in sports (Wu et al., 2007; Umeda et al., 2008), and in therapy (Pedersen et al., 1996). Despite extensive studies on manual pushing, we found that the gait cycle and laterality have received less attention from researchers. The gait cycle is described as the duration from one foot contacting the Received 7 August 2014; accepted 3 November 2014
80
KADEK HERI SANJAYA et al.
floor to the next contact of the same foot during walking (Ounpuu, 1994). It includes a stance phase and a swing phase. Laterality refers to asymmetrical preferential use of the limbs and sensory functions (Schneiders et al., 2010). Dessery et al. (2011) reported that a common mistake in ergonomic studies is that researchers focus on the dominant limb due to the assumption of human bilateral symmetry. In gait studies, the symmetry assumption is often adopted to simplify data collection and analysis (Sadeghi et al., 2000); however, this assumption means that important responses of the locomotory system are missed (Haddad et al., 2006). It is generally understood that about 85% of the human population is right-handed (RH) (Uomini, 2009). There have even been suggestions that the rate may be higher, around 90%, and that 80% of humans are right-handed (Carey et al., 2001). Population studies show that LH living in a world designed primarily for RH were found to be more accident prone (Coren, 1992). Laterality is not always constant from birth, as beside cultural influences, handedness transfer due to pathological causes has been reported (Jones, 1870). Hence, about 10% of both LH and RH were found to switch handedness (Coren, 1992). While many studies have been conducted on handedness, laterality measures such as footedness have been given less attention, despite the fact that ear preference was found to be more related to footedness than handedness (Elias et al., 1998; Chapman et al., 1987; Peters et al., 1988; Schneiders et al., 2010). No correlation between handedness and footedness has been observed (Chhibber and Singh, 1970). In particular, while RH tend to prefer the right foot, the situation remains unclear for LH (Chapman et al., 1987; Peters, 1988). In unipedal postural control, difference due to footedness was reported (Golomer and Mbongo, 2004). Dessery et al. (2011) found asymmetrical body motion influenced by footedness in gait initiation. Other studies, however, have reported an insignificant effect of footedness during walking (Zverev, 2006) and gait initiation (Hesse et al., 1997). Sadeghi et al. (2000) formulated a hypothesis that asymmetrical lower limb behaviour exists because of functional differences in propulsion and control. On the other hand, Hart and Gabbard (1997) suggested that lower limb choice for postural stabilization is taskdependent. These contradictory results have been attributed to limitations in the methods of previous studies, which mostly involved only right-lateralised subjects. Laterality has been proven to be an important feature in human life, with social, educational, and psychological implications. It affects health and well-being and even life span-LH were found to have a 2080% higher risk of injury in activities, such as sports, working, and driving, as well as shorter life expectancy (Coren, 1992). Nevertheless, the reasons for this are still not well understood. In manual pushing, laterality has received less attention. Thus, the aim of this study was to investigate the influence of laterality in various velocities of a gait cycle on manual pushing by measuring variables such as pushing force, muscular activity and foot contact duration in order to confirm whether different characteristics of LH and RH exist in manual pushing. METHODS
Subjects Seventeen healthy adult males were selected from the Chiba University student population (age 28 ± 5 years; height 169.9 ± 6.9 cm; weight 64.6 ± 7.3 kg). The subjects’ laterality was measured with the Waterloo Footedness Questionnaire (WFQ), which consisted of 10 questions with a value range from −20 to +20, and the Waterloo Handedness Questionnaire (WHQ), which contained 36 questions with a value range from −72 to +72 (Elias, et al. 1998). The questionnaire replies showed that 11 subjects were both RH (50.25 ± 12.42; positive value represents RH subject) and right-footed (9.25 ± 4.13; positive value represents right-footed subject), while 6 subjects were LH (−15.17 ± 13.7; negative value represents LH subject) and mixed-footed (0.0 ± 7.5; zero value represents mixed-footed subject). All RH were selected with a simple random sampling method; however, LH were targeted specifically from an observed LH group, since their population is very small, and consistent LH were even more difficult to find (Witelson, 1985). The number of LH (which was the maximum number obtained from the observed population) was considered sufficient, as it had been used in a previous study (Tan, 1989) and provided enough power to examine
81
LATERALITY AND ASYMMETRY IN MANUAL PUSHING
(a)
Monitor for timer display Force Plate
(b) 3rd Metatarsal
3rd Metatarsal
50 cm
8 cm
148 cm
1st Metatarsal
5th Metatarsal
5th Metatarsal Right
Left
Calcaneus
Right gait cycle
Force plate height : 105cm
Left gait cycle
Great Toe
Picture from below Picture from above
Fig. 1. Experimental set up: (a) a subject pushing a force plate on a treadmill; (b) location of foot pressure sensors attachment on the sole (left side) and gait cycle duration (right side).
statistical significant differences between groups based on physiological measurements. Prior to the experiment, the subjects were informed about the experimental procedure, and gave signed informed consent. The study was approved by the Ethics Committee of the Graduate School of Engineering, Chiba University (25-25).
Instruments As shown in Fig. 1 (a), the subjects pushed a wall force plate while walking on a motor treadmill (SportsArt 3108, SportsArt Fitness Corp., Taiwan) at three velocities (1.5, 3 and 4 km/h) in the posture they found most convenient for exerting the required pushing force. The treadmill was 194 cm (long) x 73.5 cm (wide) x 150 cm (high), with a walking lane 148 cm (long) x 50 cm (wide). Pushing force was measured with a wall force plate that contained four LMB-A-500 N load-cells (Kyowa Corp., Japan); the load-cell-to-load-cell distance was 300 mm, and all load-cells were connected to an SA-30A strain amplifier (TEAC Corp., Japan). The force plate handle was positioned 105 cm above the ground, a height considered capable of accommodating a range of subjects with various body heights (Lee et al., 1991; Resnick and Chaffin, 1995). EMG data were collected bilaterally from the following muscles: tibialis anterior (TA), soleus (Sol), lumbar erector spinae (L5ES), and triceps brachii (Triceps). Both the strain amplifier output and EMG electrodes were connected to an MP 150 data acquisition system (Biopac Systems, USA), whose output was then connected to a personal computer. All trials were recorded using two digital cameras (Canon Ixy, Canon Corp., Japan), placed on the left side of and behind the subjects, to confirm visually post-experiment that no awkward posture had been performed during the experiment. If such a posture was observed, that data was removed from further analyses. To simplify the visual confirmation, ball markers were attached on the subjects’ acromion, elbow, greater trochanter, knee and lateral malleolus. In this study, however, we did not perform kinematic analysis. As shown in Fig. 1 (b), to measure the timing of the sole contacting the ground, ten FSR-400 pressure sensors (Interlink Electronics, USA) were attached bilaterally on the great toe (left: LT, right: RT), 1st metatarsal (left: L1MT, right: R1MT), 3rd metatarsal (left: L3MT, right: R3MT) and 5th metatarsal (left: L5MT, right: R5MT), and the calcaneus (left: LC, right: RC) based on the method used by Kiriyama et al. (2005). All the subjects wore the same footwear. Trial duration was marked by a visual display timer and a light sensor (Kodenshi Corp., Japan). Experimental Procedures The subjects performed pushing trials while walking on a treadmill at three velocities: 1.5 km/h (P1.5), 3 km/h (P3) and 4 km/h (P4) (Fig. 1a). The slope angle of the treadmill was set at 0º (flat surface). The experiment was performed inside a climate chamber with controlled temperature (26ºC) and lighting. During
82
KADEK HERI SANJAYA et al.
the main trials, the subjects were instructed to push at about 50% of their maximum static pushing force. For each condition, trials were performed in 10-s intervals, three times, with a three-min rest between trials. From the 10-s measured data, three gait cycles in the middle, which is considered the most stable period, were processed for further analyses. All trials were randomized. During each trial, the subjects were instructed to gaze at the monitor that showed both a timer and pushing force feedback. This gazing was helpful in avoiding the asymmetrical visual influence on balance control as reported in previous studies (Bessou et al., 1999; Golomer and Mbongo, 2004; Nagano et al., 2006). One gait cycle
(a)
RGT (right great toe) pressure shows: pressure magnitude and duration
Volt
RGT 10 0
Volt
R1MT 10 0
Volt
R3MT 10 0
Volt
R5MT 10 0
Volt
RCalc 10 0
0
(b)
500
1000
10
1500
2000 time(ms)
2500
3500
4000
Swing Phase
Stance Phase
9
3000
8 7
Volt
6
RGT R1MT
5
R3MT
4
R5MT
3
RCalc
2 1 0
0
20
40
60
80
100
time (normalized into 100 data points)
Fig. 2. Data analysis of gait cycle duration from raw data transformed using AcqKnowledge software into a processed graph in Microsoft excel. (a) a typical 4s recording of pressure sensors attached on the right sole. (b) data taken from the five pressure sensors during one gait cycle, which has been normalized into 100 data points, showed swing phase and stance phase. RGT: right great toe; R1MT: right 1st metatarsal; R3MT: right 3rd metatarsal; R5MT: right 5th metatarsal; RCalc: right calcaneus.
Data Analysis As shown in Fig. 1 (b), data analysis was based on one gait cycle, from one heel-strike to the next heel-strike of the same foot, marked by the calcaneus foot pressure sensor. Fig. 2 shows the raw data transformed by using AcqKnowledge software (Biopac systems, USA) into Microsoft excel (Microsoft Corp., USA). The stance phase duration of one gait cycle was a combination of the five pressure sensors attached on the sole (Fig. 2). Each pressure sensor showed pressure magnitude and duration of a sole point. We di-
83
LATERALITY AND ASYMMETRY IN MANUAL PUSHING
vided data based on right- and left-side gait cycles, which meant that data from left-side muscles and pressure sensors were analysed for left foot (LF) gait cycles, and data from the right side were analysed according to right foot (RF) gait cycles. Force plate, EMG and pressure sensor data were collected at a 1000-Hz sampling rate. Raw EMG signals were band-pass filtered between 15−250 Hz and root mean square (RMS) was derived. All data were normalized into 100 data points during one gait cycle. To measure temporal changes in symmetry of variables measured on the left and right sides, the crosscorrelation function (CCF) was used. The coefficients (Rxy) vary between −1 and +1, where a positive correlation value indicates the time-varying signals that are in phase or increasing and decreasing together, and a negative value indicates an inverse relationship (Nelson-Wong et al., 2009). CCF also measured time lag (τ) between two signals. Data measured from the left side during a LF gait cycle were used as template and data from the right side during a RF gait cycle were used as data. If the peak of the right side data occurred earlier than that on the left side, time lag was considered positive; if it were later, time lag was considered negative. In further statistical analyses, the subjects were grouped into RH and LH groups, since we wanted to compare RH and right-footed subjects who were majority of the population with other laterality groups. We analyzed the characteristics of each group, and for CCF and time-lag data, we also performed comparisons between groups. The Shapiro-Wilk test was used to determine the probability normal distribution of the data. To compare data between left and right gait cycles, we employed Student`s paired t-test for parametric method and Wilcoxon signed-rank test for non-parametric method. Data comparison between different velocities in the same group was performed using one-way repeated measures ANOVA with Bonferroni post-hoc test for parametric method and Friedman Test with the Wilcoxon signed-rank posthoc test for non-parametric method. Because the subjects groups had unequal samples, in order to compare groups, independent t-test for parametric method and Mann-Whitney U test for non-parametric method were used. Statistical significance was set at p
