Assessing Contextual Interference as a Learning Tool for Gait Modifications
Introduction
Obesity is currently a serious health issue in our country, as over one third of the U.S. population is considered obese with a prevalence rate among adults of approximately 35.6% (Flegal, Carroll, Kit, & Ogden, 2012). Obesity can lead to a myriad of other diseases including diabetes, high blood pressure, cardiovascular disease (CVD), or osteoarthritis (OA), just to name a few. OA is a degenerative disease that occurs when the joint cartilage becomes worn down and lost over time, resulting in increased pain and loss of mobility in the affected joint. Knee OA, specifically, is the most common joint affected by OA and can be found in 10% of men and 13% of women over the age of 60 (Freedman Silvernail, Milner, Thompson, Zhang, & Zhao, 2013). This loss of cartilage often leads to an increase in knee pain further inhibiting exercise, active lifestyles, and ability to perform daily tasks, which compounds the obesity problem Mendes et al., (2018). Gait research is specifically paramount as walking is the most common form of physical activity world-wide and is also the most common prescribed form of exercises specifically in obese populations (Flegal et al., 2012). Due to the severity of the obesity epidemic, there has been a flurry of research to determine obesity’s effect on gait biomechanics and explore different strategies to reduce abnormal joint loading during gait.
For example, research has found that obese individuals often present more extreme gait biomechanics including increased ground reaction forces and joint contact forces compared to healthy weight individuals, while also walking at slower speeds (de Souza et al., 2005). It is suspected that obese individuals often use this slower walking speed as a protective mechanism to reduce vertical ground reaction forces (VGRF). Several studies have also found that obese individuals present increased knee abduction and extension moments during level walking compared to healthy weight individuals (Blazek, Asay, Erhart-Hledik, & Andriacchi, 2013; Yocum, Weinhandl, Fairbrother, & Zhang, 2018,). This is important to consider as knee joint moments are the most common indicator of knee joint loading and development of knee osteoarthritis (Paquette, Zhang, Milner, Fairbrother, & Reinbolt). It is also hypothesized that this reduction in walking speed is a strategy for reducing detrimental or abnormal peak knee abduction moments, which can be indicative of knee abnormal joint loading. (Freedman Silvernail et al.).
In biomechanics, an up and coming area of research is how different gait modifications can be implemented to improve joint kinematics and kinetics at the hip, knee or ankle. Some examples of different gait modifications can include foot progression angle, toe-in/toe-out gait, or step width. There have been several approaches to implementing various gait modification to reduce knee joint moments such as abduction and external rotation moments. When specifically focusing on knee joint biomechanics, these can be useful, nonsurgical strategies for reducing abnormal knee joint loading (Fregly, Reinbolt, Rooney, Mitchell, & Chmielewski, 2007 J. A., Rooney, K. L., Mitchell, K. H., Chmielewski, T. L., 2007), and step width has been proven as an effective gait modification to reduce knee abduction moments in healthy-weight and obese populations (Yocum et al., 2018) (Paquette et al., 2014) (Bennett, Shen, Cates, & Zhang, 2017) (Brindle, Milner, Zhang, & Fitzhugh, 2014). This is important as joint moments can be indicative of loading at the respective joints (Freedman Silvernail et al., 2013; Paquette et al., 2014).
Step width as a gait modification is measured in the frontal plane, and can be predicted by an individual’s leg length. Specifically, in a study conducted by Donelan et al., in 2001 (Donelan, Kram, & Kuo, 2001) that an individuals preferred step width is approximately 13% of their leg length. In a following study, it was determined that step widths of 26% and 39% of the individual’s leg length are significantly different from their preferred width to result in different walking biomechanics (Paquette et al., 2014).
During level walking, a study conducted by Yocum et al. (2018) reported that as step width increased, knee abduction moment was significantly decreased in obese populations. Specifically, it was reported by Yocum et al. (2018 10) that increasing step width significantly reduced peak knee adduction angle at loading response in both healthy weight and obese individuals. A different study, concerned with the effects of increased step width during level running, conducted by Brindle, R. A., Milner, C. E., Zhang, S., Fitzhugh, E. C. (2014 46) also found that at the wider step width, peak knee abduction moment and knee abduction impulse were significantly reduced. These results supported those found by Yocum et al. (2018) that as step width was increased, knee abduction moment was significantly decreased in obese populations.
When testing gait modifications in the lab, targets are used to ensure that the participant is able to adhere to the gait modification however, this does not necessarily aid in motor learning of that new modification and it is unknown how different gait modifications is transferred to daily living. In order to investigate whether newly learned gait modifications can be learned and adhered to outside of the lab setting I propose using the principle of contextual interference (CI) to test how well a new gait modification is learned in a motor learning context.
In motor behavior literature, CI has been shown as a successful strategy to increase retention and learning of a new motor skill. Studies have shown that moderate levels of CI can be more beneficial than no CI at all, and while practice will yield lower scores, individuals present superior retention of a motor skill compared to no CI.
The purpose of this study is to test CI as a method for learning new gait modifications in a laboratory setting and being able to apply them in a natural walking environment. I hypothesize that moderate levels of CI will be beneficial as a motor learning strategy for step width as a gait modification. Specifically, those individuals exposed to a higher level of CI will have a wider average step width and as a result, greater reduction of knee abduction and extension moments compared to lower CI during the transfer test. I also hypothesize that the step width data from the practice trials will be less consistent for the higher CI group compared to the low CI group. A transfer test is critical to use in order to understand how in lab gait modifications can be applied outside the lab.
Methods and Materials
Participants
For this study, twenty participants will be recruited through flyers posted in building on the University of Tennessee campus and word of mouth. Those who are recruited and meet the criteria for inclusion and exclusion will be asked to participate and sign an informed consent form approved by The University of Tennessee Institutional Review Board (IRB). The participants will then be randomly separated in to two groups of ten participants each, a lower CI group and a higher CI group.
To be included in the study, participants must be 18 – 30 years old. Any individuals who cannot walk without assistance, have been diagnosed with any joint disease, have had a previous surgery or major injury (fracture, meniscal/ligament tear) in the lower extremity one year prior to participation or a BMI value more than 38kg/m2 will be excluded. This BMI value was selected as a cut-off due to the difficulty of accurately placing and tracking skin mounted markers.
Instrumentation
For three-dimensional (3D) motion data collection during the test trials, a twelve-camera motion analysis system (240 Hz, Vicon Motion Analysis Inc., Oxford, UK) will be used. To collect kinematic and kinetic data, an instrumented treadmill (AMTI, 176 Waltham Street, Watertown, MA 02472-4800 USA) will be used (Figure 1). All data will be exported to a biomechanical analysis software suite (Visual3D, 2.6, C-Motion, Inc., Germantown, MD, USA) and preferred step widths will be calculated by finding the mediolateral distance between the center of masses of both feet during their respective midstance in the analysis software suite.
During the transfer test, participants will walk through a 10-meter-long hallway which has a walking pathway of 5 meters in length with AMTI force platforms (1200 Hz, BP600600 and OR-6-7, American Mechanical Technology Inc., Watertown, MA, USA) (figure 1). This walking pathway will have a 12 camera motion analysis system (240 Hz, Vicon Motion Analysis Inc., Oxford, UK) such that one camera is located on each end and the other ten cameras are spaced evenly along either side of the walking path. To measure and control for walking speed, two photocells (63501 IR, Lafayette Instrument Inc., IN, USA) and two electronic timers (54035A, Lafayette Instrument Inc., IN, USA) will be used. This will ensure that a participants walking speed in the transfer test is not significantly different from the speeds they chose to walk at on the treadmill. This is important as any significant changes in speeds will affect joint moments.
For all trials, participants will wear spandex shorts provided by either the lab or themselves, a tight fitting shirt, and standard lab shoes (Nike Pegasus 3.0). Reflective markers will be placed on bony landmarks bilaterally on the participant’s acromion process, iliac crest, greater trochanter, medial femoral epicondyle, lateral femoral epicondyle, medial malleolus, lateral malleolus, 1st metatarsal head, 5th metatarsal heads and the 2nd toe. These landmarks serve as the anatomical landmarks needed during the static calibration trials. For the tracking markers, four of the retroreflective markers attached to a thermoplastic plate will be placed on the posterior trunk, posterior aspect of the pelvis (two- marker cluster on each side), lateral surface of thighs and shanks, and four markers will be placed on the heel cup of each shoe.
Figure 1. An example of the instrumented treadmill used in the practice trials (left) and the walkway used in the transfer test (right).
Experimental Procedures
Pre-test: Prior to data collection, the participant will be provided with the informed consent form and complete a participant information form in which demographic information such as height and weight to calculate BMI value as well as shoe size, etc. The participant will also fill out a survey, Knee Injury and Osteoarthritis Outcome Score (KOOS), about knee functions and finally a PAR-Q+ survey to determine the activity levels of participants or if any participants have had any major surgery or injury in the lower extremity in the past 6 months that would result in their exclusion from the study.
Participants will then complete a short, 5-minute warm-up at a self-selected speed on the treadmill to become accustomed to walking on the treadmill as well as wearing the lab provided footwear. Following this warmup, retroreflective markers will be attached as previously mentioned, and measurements of leg length and shoulder height will be taken. In this study, leg length will be measured as the distance from the anterior superior iliac spine to the lateral malleolus while the participant is lying down, as was done in the study conducted by Brindle et al., (2014). These measurements are needed to determine the increased step widths for each participant during the practice trials and the appropriate height for the photocells during the post-test.
Once the markers are attached to the participant, a static trial will be taken after which the anatomical markers will be removed. The participant will then complete a 3-minute walking trial to collect preliminary kinetic and kinematic data such as their preferred step widths. In order to calculate wide step width for each participant, the average step width from the pre-test walking trial will be calculated and 13% of their leg length will be added on top of this value, resulting in an approximate width of 26% of the participant’s leg length. Wider step width will be calculated in a similar fashion, by adding another 13% onto the WSW for a total of approximately 39% leg length. These percentages were selected from previous research by (Brindle et al., 2014) who determined that these widths were significantly different from each other as mentioned previously.
Practice trials: Over the course of five days, participants will come in for approximately 90 minutes each day for data collections. Before each test session, the participant will be asked to complete a five-minute warm-up and will have static trials taken as it is impossible to ensure marker placement will be the same each day. Both groups will complete a total of 15 walking trials, five for each of the three step width conditions lasting three minutes each at their preferred walking speeds. Each participant’s preferred speeds will be recorded to use during the transfer tests.
Participants assigned to the low CI group will perform their trials in a blocked fashion, three blocks with five trials each, increasing step width in each block. They will be allowed to rest for 2-3 minutes between each block of trials. The higher CI group will perform their 15 trials in a serial fashion, completing five rounds of three trials, 1 trial at preferred step width, 1 trial at wide step width, 1 trial at wider step width. The higher CI group will also be allowed 2-3 minutes of rest between each round. For both groups, during the wide and wider step width conditions, there will be a mirror placed in front on the participant, with lines of tape indicating the target step width. This will enable the participant to line their feet up with the tape to achieve the target step widths. This mirror will not be present for either group during the preferred step width conditions.
Transfer test: While the pre-test and practice trials will occur on a treadmill for convenience in providing feedback about step width and monitoring the participant’s performance, the transfer tests will take place on a stationary walkway to simulate walking conditions experienced in day to day living (walking in an office building, through a grocery store aisle, etc.) The path of force plates will be 5 meters long and 1.5 meters wide (figure 1) to ensure the participant is not constrained by the size of the force plates. The participants from each group will be instructed to walk freely across the walkway at their preferred walking speed. Walking speeds will be monitored by the photocells and timing gates mentioned previously. Any trials that are greater than 10% different from the preferred speed used during the practice trials will be discarded and the participant will be asked to repeat the trial with feedback informing them if they were too fast or too slow.
Data analysis
3D kinematic trajectories will be analyzed in the Nexus (2.3, Vicon Motion Analysis Inc., Oxford, UK) to ensure that all markers are correctly labelled, no gaps are present in the data, and any ghost markers are removed from the data. Once this is completed, all data will be imported into a 3D data analysis software suite, Visual3D (version 2.6, C-Motion, Inc., Germantown, MD, USA) for 3D kinematic and kinetic analysis. For kinematic analysis, all data will be computed with an X-Y-Z Cardan rotational sequence where positive values are indicative of extension and adduction and ankle dorsiflexion and inversion angles in accordance with the right hand rule. Kinematic data will be filtered with a zero-lag fourth-order Butterworth low-pass filters at a cutoff frequency of 6Hz, while GRF data are filtered at a cutoff frequency of 50Hz, to clean the data.
Statistical Analysis
Means and standard deviations for step width, vertical ground reaction force, mediolateral ground reaction force, 1st and 2nd peak knee extensor moments, 1st and 2nd peak knee abduction moments will be collected for each group from all practice and transfer trials. The dependent variable(s) (DV) being examined are: step width, 1st and 2nd peak knee extensor moments, 1st and 2nd peak knee abduction moments. The independent variable (IV) is level of contextual interference.
A 2 × 3 (CI group X Step Width) two-way mixed designed ANOVA will be used to determine how differences in contextual interference affect step width, peak knee joint kinematics, moments, and ground reaction forces in a transfer test. Significant results will be defined as having a p value ≤ 0.05.
Limitations
There are a few limitations to this study, the first being the health status of participants. Individuals suffering from various injuries or diseases may not be able to adopt a gait modification as easily as a healthy population. Another limitation is that skin marker placement may not accurately reflect bony landmark locations, to combat this any participant with a BMI value < 38 kg/m2 will be excluded as higher BMI levels decrease tracking accuracy of skin mounted markers due to increased adipose tissue. Reflective markers used to track the feet will be placed on the shoe, and therefore may not accurately reflect the motion of the foot within the shoe. Finally, the transfer test will be conducted in a laboratory setting free from walking hazards such as uneven ground or other distractions.
References:
Bennett, H. J., Shen, G., Cates, H. E., & Zhang, S. (2017). Effects of toe-in and toe-in with wider step width on level walking knee biomechanics in varus, valgus, and neutral knee alignments. Knee, 24(6), 1326-1334. doi:10.1016/j.knee.2017.08.058
Blazek, K., Asay, J. L., Erhart-Hledik, J., & Andriacchi, T. (2013). Adduction moment increases with age in healthy obese individuals. J Orthop Res, 31(9), 1414-1422. doi:10.1002/jor.22390
Brindle, R. A., Milner, C. E., Zhang, S., & Fitzhugh, E. C. (2014). Changing step width alters lower extremity biomechanics during running. Gait Posture, 39(1), 124-128. doi:10.1016/j.gaitpost.2013.06.010
de Souza, S. A., Faintuch, J., Valezi, A. C., Sant' Anna, A. F., Gama-Rodrigues, J. J., de Batista Fonseca, I. C., . . . Senhorini, R. C. (2005). Gait cinematic analysis in morbidly obese patients. Obes Surg, 15(9), 1238-1242. doi:10.1381/096089205774512627
Donelan, J. M., Kram, R., & Kuo, A. D. (2001). Mechanical and metabolic determinants of the preferred step width in human walking. Proc Biol Sci, 268(1480), 1985-1992. doi:10.1098/rspb.2001.1761
Flegal, K. M., Carroll, M. D., Kit, B. K., & Ogden, C. L. (2012). Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA, 307(5), 491-497. doi:10.1001/jama.2012.39
Freedman Silvernail, J., Milner, C. E., Thompson, D., Zhang, S., & Zhao, X. (2013). The influence of body mass index and velocity on knee biomechanics during walking. Gait Posture, 37(4), 575-579. doi:10.1016/j.gaitpost.2012.09.016
Fregly, B. J., Reinbolt, J. A., Rooney, K. L., Mitchell, K. H., & Chmielewski, T. L. (2007). Design of patient-specific gait modifications for knee osteoarthritis rehabilitation. IEEE Trans Biomed Eng, 54(9), 1687-1695. doi:10.1109/TBME.2007.907637
Mendes, J., Borges, N., Santos, A., Padrao, P., Moreira, P., Afonso, C., . . . Amaral, T. F. (2018). Nutritional status and gait speed in a nationwide population-based sample of older adults. Sci Rep, 8(1), 4227. doi:10.1038/s41598-018-22584-3
Paquette, M. R., Zhang, S., Milner, C. E., Fairbrother, J. T., & Reinbolt, J. A. (2014). Effects of increased step width on frontal plane knee biomechanics in healthy older adults during stair descent. Knee, 21(4), 821-826. doi:10.1016/j.knee.2014.03.006
Yocum, D., Weinhandl, J. T., Fairbrother, J. T., & Zhang, S. (2018). Wide step width reduces knee abduction moment of obese adults during stair negotiation. J Biomech, 75, 138-146. doi:10.1016/j.jbiomech.2018.05.002