Pedaling Asymmetries in Cyclists With Unilateral Transtibial Amputation: Effect of Prosthetic Foot Stiffness

Transcription

1 Journal of Applied Biomechanics, 2011, 27, Human Kinetics, Inc. Pedaling Asymmetries in Cyclists With Unilateral Transtibial Amputation: Effect of Prosthetic Foot Stiffness W. Lee Childers, Robert S. Kistenberg, and Robert J. Gregor Georgia Institute of Technology Cyclists with unilateral transtibial amputation (CTA) provide a unique model to study integration of the neuromuscular and bicycle systems while having the option to modify this integration via the properties of the prosthesis. This study included eight CTA and nine intact cyclists. The cyclists pedaled on a stationary bicycle with instrumented force pedals. The CTA group pedaled with a stiff or flexible prosthetic foot during a simulated time trial and a low difficulty condition. During the time trial condition, pedaling with the flexible foot resulted in force and work asymmetries of 11.4% and 30.5%, the stiff foot displayed 11.1% and 21.7%, and the intact group displayed 4.3% and 4.2%, respectively. Similar trends were shown in the low difficulty condition. These data suggest foot stiffness has an effect on cycling symmetry in amputees. Keywords: amputation, motor control, cycling, prosthetic design Symmetry in human locomotion is often assumed whereas asymmetry seems to be the more apparent outcome, even in the absence of pathology (Sadeghi et al., 2000, for review). Asymmetry during gait has been documented in persons with amputation (Winter & Sienko, 1988; Silverman et al., 2008), but difficulty separating the balance and propulsion aspects of gait prevents a clear understanding of how the prosthesis is used for propulsion. Cycling, as a form of locomotion, requires effective integration of the human and bicycle systems, while the coupled nature of this task provides a unique environment for the investigation of symmetry; that is, it allows for energy transfer between limbs while the upper body remains supported by the saddle, limiting the need for balance. Asymmetries in cycling have been reported in intact populations (Daly & Cavanagh, 1976; Sanderson, 1990; Smak et al., 1999), indicating asymmetry exists but is relatively low, such as 10%, and varies with pedaling resistance. Pedaling asymmetry may also develop, for example, when pathology creates deficiencies in the ability to generate and direct force (Perell et al., 1998; Hunt et al., 2004). Furthermore, pedaling asymmetries have been quantified with regard to force (related to limb strength and inertial properties) and work (related to control of force direction) production (Sanderson, 1990). Hence, measuring force and work asymmetries during cycling may expose deficiencies and compensation strategies in limb control. W. Lee Childers (Corresponding Author), Robert S. Kistenberg, and Robert J. Gregor are with the School of Applied Physiology, Georgia Institute of Technology, Atlanta, GA. Cyclists with unilateral transtibial amputation (CTA) provide a unique model to study integration of the neuromuscular and bicycle systems. These cyclists must now adapt to the structural and physiological changes related to the amputation and interact with their environment through a prosthetic limb on one side and an intact limb on the other. Asymmetry in three Paralymic cyclists with amputation has been reported in popular literature as being between 29.2 and 39.8% (Broker & Gregor, 1996). The authors expressed caution in generalizing these data, however, because of the small sample size three subjects and the elite fitness level in the cyclists. There are no reports providing insight into how lesser trained cyclists may compensate for the loss of a limb. An interesting aspect of the case of a CTA is the potential to modify and perhaps optimize the human-cycling system through prosthetic foot selection. Ankle joint stiffness during cycling has been shown to affect pedaling asymmetry (Pierson-Carey et al., 1997). During intact cycling, the soleus is active during the power phase while the tibialis anterior is active during recovery to stabilize the ankle joint allowing energy generated by more proximal muscles to be transferred to the crank (Ryan & Gregor, 1992; Kautz & Neptune, 2002). Whereas ankle joint stiffness may be actively controlled by an intact neuromuscular system, ankle joint stiffness of a prosthetic limb is set by the passive mechanical properties of the prosthetic foot (Hafner et al., 2002). Therefore, altering stiffness of the prosthetic foot could impact energy transfer from remaining muscles in the amputated limb, thus providing a method to perturb the bicycle/rider system in a CTA. The purpose of this project was to quantify the amount of pedaling asymmetry in a CTA group compared with that of a group of intact cyclists, as well as 314

2 Cyclists With Unilateral Transtibial Amputation 315 to investigate the effect of prosthetic foot stiffness. Our hypotheses were that (1) asymmetry in cyclists with unilateral amputation will be greater than that of intact cyclists with regard to force and work output, (2) work asymmetry will be greater than force asymmetry in CTA, and (3) asymmetry will decrease as prosthetic foot stiffness increases. In addition, knowledge gained in this study may provide evidence-based recommendations for cycling-specific prosthetic design. Currently, prosthetic foot selection for cycling has no established protocol and has been left to the opinion of health care professionals. Typically, prosthetic feet designed for walking have been adapted for use in cycling with no systematic evaluation of their performance (Gailey, 2004). Methods Subjects and Protocol Eight CTA (six males and two females; average age 36.4 ± 10.4 years, height 1.75 ± 0.07 m, and mass ± 10.8 kg) and a control group of nine intact cyclists (eight males and one female; average age 30.4 ± 13.4 years, height 1.81 ± 0.06 m, and mass 74.0 ± 6.6 kg) volunteered to participate in the study. Cycling experience was generally recreational for both groups. However, one cyclist in the CTA group was a Paralympic medalist and three cyclists in the intact group held amateur racing licenses. Informed consent was obtained from all subjects before the beginning of data collection. The study protocol and informed consent forms were approved by the institutional review board. In addition to the anthropometric measurements, thigh circumferences were taken midway between the greater trochanter and the lateral epicondyle of the femur on each leg to indirectly estimate differences in muscle mass between limbs. Inclusion criteria for both groups were that subjects had to ride a bicycle at least once a month and be between 18 and 65 years of age. Inclusion criteria specific to the CTA group included having a unilateral transtibial amputation for at least 1 year, with the cause of amputation being traumatic, congenital, or tumor. Exclusion criteria for both groups included cardiovascular or pulmonary pathologies, muscle paralysis, and open skin sores on the lower limbs. The subjects rode either their personal bicycle or a bicycle adjusted to the saddle and handlebar location of their personal bicycle. The bicycles were mounted in a stationary trainer with a calibrated centripetal resistance unit (1-UP USA Inc.). All bicycles were adapted with 172-mm crank arms and piezoelectric element force pedals (Broker & Gregor, 1990). The rationale behind the crank arm length selection was that this length was between the shortest (170 mm) and longest (175 mm) crank lengths of the subjects personal bicycles. The crank arms used in this study were machined to fit a gear that would a drive rotary potentiometer providing crank position and, as such, did not allow for the subject s specific crank length. Pedal/rider interface was either through a Look (Look Corp.) or Shimano SPD (Shimano Corp.) pedal system based on subject preference. Interfaces were adapted to fit the piezoelectric element force pedals (Wheeler et al., 1992). Both interfaces would allow a total of 8 of foot axial rotation relative to the pedal (± 4 from neutral) but would not allow translation in any other direction between the foot and pedal. Prosthesis protocol for the CTA group was performed by a certified prosthetist. The subject s personal prosthesis was modified by removing the pylon and foot section and replacing it with an aluminum pylon with a stiff aluminum foot (STIFF condition) section or a flexible carbon-fiber dynamic response prosthetic foot (FLEX condition) (Figure 1). The stiffness of the prosthetic foot was the most flexible foot that the manufacturer (Ossur Corp.) would allow for safe ambulation and was determined based on subject body mass. The cycling prostheses retained the subject s socket, suspension method, and alignment. An Otto-Bock laser-assisted posture device (Otto Bock Inc.) was used to ensure alignment consistency. The prosthetic foot was aligned with minimal toe-out to avoid a clearance problem between the posterior aspect of the foot and the crank arm. The subjects wore a cycling shoe with a stiff sole on their sound side. The cycling cleat was placed in the center of the foot in the coronal plane and aligned with the subject s first metatarsal head on his or her sound limb in the sagittal plane for both feet. Load conditions appeared in random order after a 5-min warm-up and included a resistance and cadence combination selected by each subject defined as low difficulty (LD) and a 6-min simulated time trial (TT). These relative load conditions were used to better replicate the cycling task in a laboratory setting (Carpes et al., 2007). Each load condition lasted for 6 min, during which data were collected during the last minute. A heart rate monitor was worn during all trials, and heart rate recorded once during the final minute of each load condition. Workload difficulty was defined as the percent of the subject s maximum age-predicted heart rate: 220 age of the subject (Fox et al., 1971). A centripetal resistance unit was used with the stationary trainer. This design varied resistance via changes in rear wheel speed. Maintaining a constant rear wheel speed then ensured constant power output between trials. The rear wheel speed was displayed to the subjects via a cyclecomputer (Sigma Corp.). Subjects could vary the load by shifting to a different gear and by varying their cadence but were required to maintain a constant rear wheel speed and gear selection once the LD and TT conditions were determined. The subjects actual power output was calculated from instrumented force pedals described below. Kinetic and Prosthetic Foot Deflection Analysis The normal and tangential components of the pedal reaction force were measured using dual piezoelectric

3 316 Childers, Kistenberg, and Gregor Figure 1 Cycling prosthesis used. The stiff aluminum foot is on the left and the flexible carbon-fiber foot is on the right with the linear potentiometer mounted to measure foot deflection. The prosthesis retained the subjects prosthetic socket but replaced the shank and foot section with a stiff aluminum piece to maintain consistency across subjects. instrumented force pedals (Broker & Gregor, 1990). A linear potentiometer was used to measure prosthetic foot deflection of the FLEX foot. This was affixed to the pylon at one end via a bracket and the bolt holding on the cycling cleat at the opposite end (Figure 1). The sensor was aligned to measure displacement normal to the pedal surface. The linear position sensor was not used during the STIFF foot trials because the calculated maximum foot deflection of the aluminum plate used would be less than 0.16 mm. This small amount of deflection was within the measurement error of the linear position sensor. Pilot testing confirmed that the STIFF foot did not have any measurable foot deflection and thus assumed to be zero. Data were recorded at 300 Hz, and a moving average with a three-sample window was subsequently used to smooth the force data. Time of each cycle was normalized to 100 data points, and five complete pedal cycles from each trial were averaged together for analysis. Position of the crank and pedal orientation were measured with gear-driven continuous turn potentiometers. The top dead center of the right pedal was defined as the crank zero position. The energy released and absorbed by the prosthetic foot was calculated by multiplying the force orthogonal to the pedal surface and the foot deflection. The ratio of energy released to energy absorbed was calculated by dividing the negative impulse by the positive impulse. Asymmetry Pedal force data were reduced into components about the longitudinal axis of the crank and an axis perpendicular to the crank. Torque about the crank spindle was calculated for each pedal by multiplying the perpendicular force by the crank arm length. Total work done to rotate the crank by each limb was then determined by integrating the crank torque curve. The dominant limb was defined as the limb contributing the most work to the average pedal output. Work asymmetry was calculated as the difference between the percentage of work produced by the dominant limb and the same percentage was calculated for the nondominant limb. For example, if the dominant limb accounted for 55% of the total work, then the nondominant limb accounted for 45%, and work asymmetry would be 10%. These steps were repeated to find force asymmetry as well, which was calculated as the difference between the total linear impulses for both limbs (Sanderson, 1990). Statistical Analysis A power analysis was conducted using G-power software (Faul et al., 2007) to determine the number of subjects necessary to compare the CTA and intact groups. This was performed using data published on CTA (Broker & Gregor, 1996) and intact subjects cycling at 235 W and 80 rpm (Sanderson, 1990). The results of this analysis indicated a minimum number of subjects needed to show difference between asymmetries in intact and CTA groups was 4 to achieve a statistical power of Variables describing the load conditions were compared between the CTA and intact groups to determine if both groups performed at the same relative difficulty

4 Cyclists With Unilateral Transtibial Amputation 317 and to verify that the task difficulty and torque TT and LD conditions were different from each other. A two-tailed paired t test was used to test significance between dominant and nondominant limbs for each load condition, within the intact group as well as the STIFF and FLEX foot conditions within the CTA group. A two-tailed paired t test was also used to test within the CTA group between the STIFF and FLEX foot conditions for the TT and LD load conditions. Task difficulty, torque, and cadence as well as force and work asymmetry were compared using a one-way ANOVA with four levels (load condition, intact group, STIFF foot CTA group, FLEX foot CTA group). When significant F-values existed, a Tukey post hoc was used to determine significance between particular groups. Significance was defined as a p-value 0.05 for all statistical analyses. Results There were statistically significant differences between the TT and LD load conditions in percent maximum heart rate and torque. All groups pedaled at a similar cadence for both load conditions (Figure 2). There were no statistically significant differences between foot conditions within the CTA and intact groups within a load condition regarding percent maximum heart rate, suggesting that the difficulty of the task was matched to the subject s own capabilities. The intact group did pedal at a higher torque than the CTA group (both foot conditions). The only anthropometric variable that was statistically significant between groups was the percent difference in thigh circumference between the two legs (6.1 ± 1.8% for CTA vs. 1.4 ± 0.9% for intact). The difference in contribution between the dominant and nondominant limb to total work was significantly different in both groups in both conditions (Table 1). Figure 2 Group means for task difficulty expressed as a percentage of the subjects maximum heart rate, Cadence in rev/min and torque in N m ± SD. The CTA cycling with the stiff foot (black bars), the CTA cycling with the flexible foot (hatched bars), and the intact group (white bars) for both the time trial and low difficulty conditions. *Indicates significant difference from the intact group. Indicates significant difference from the time trial condition. Table 1 Percent contribution of each limb to total work and force; mean ± SD Contribution to Total Work (%) Contribution to Total Force (%) Dominant Limb Nondominant Dominant Limb Nondominant Time Trial STIFF 60.7 ± ± ± ± 4.2 FLEX 64.6 ± ± ± ± 3.4 Intact 52.1 ± ± ± ± 1.5 Low Difficulty STIFF 71.6 ± ± ± ± 3.5 FLEX 67.5 ± ± ± ± 4.1 Intact 53.8 ± ± ± ± 2.3 Indicates significant difference from the nondominant limb.

5 318 Childers, Kistenberg, and Gregor The CTA group pedaled with significantly more work and force asymmetry than did the intact group in both LD and TT conditions (Table 2 and Figure 3). The work asymmetry was significantly greater than the force asymmetry in the CTA group for both LD and TT conditions. The dominant limb was always the sound limb for both force and work asymmetries. The difference in torque output between limbs can be seen graphically in the example data shown in Figure 4. Table 2 Force asymmetry and work asymmetry results; mean ± SD Work Asymmetry (%) Force Asymmetry (%) Time Trial STIFF 21.4 ± 8.1 * 11.2 ± 8.3 * FLEX 29.2 ± 8.0 * 11.4 ± 6.8 * Intact 4.2 ± ± 3.0 Low Difficulty STIFF 43.3 ± 23.7 * 12.0 ± 7.0 * FLEX 35.1 ± 25.3 * 11.4 ± 5.5 * Intact 7.9 ± ± 3.7 *Denotes statistical significance from the intact group. Denotes statistical significance from the FLEX foot condition within the group. Figure 3 Group means for percent work and force asymmetries ± SD during the time trial condition (top) and the low difficulty condition (bottom). The CTA cycling with the stiff foot (black bars), the CTA cycling with the flexible foot (hatched bars), and the intact group (white bars) for both the time trial and low difficulty conditions. *Indicates significant difference from the intact group. Indicates significant difference between prosthetic feet. Figure 4 Group mean torque for the dominant and nondominant limbs in the intact group (top) and sound and amputated limbs in the CTA group (bottom). Shaded area indicates ± 1 SD. Crank position starts with the right pedal at TDC. Asymmetry can be seen by comparing the right and left limbs.

6 Cyclists With Unilateral Transtibial Amputation 319 The average foot deflection of the FLEX foot during the TT condition was 3.26 mm ± 2.21 mm. Foot deflection during the LD condition was 1.36 mm ± 1.57 mm. Foot deflections normalized to the maximum within a subject and averaged across the group are illustrated in Figure 5. The prosthetic foot was compressing (storing energy) during the first 180 of the crank cycle and then decompressing (releasing energy) during the last 180. The average ratio of energy released by the FLEX foot to the energy absorbed by the prosthetic foot was 0.83 ± 0.04 during the TT condition. This value is significantly less than 1, assuming a 95% confidence interval, standard deviation of 0.04, and a sample size of 8. Discussion Two types of asymmetries were calculated during this experiment. The first, force asymmetry, represents the end result of all forces developed to load the crank, regardless of orientation. This value represents the net muscular effort with a difference between limbs indicating a difference in strength, inertial and/or anatomical characteristics (Sanderson, 1990). The second, work asymmetry, accounts for the relative contribution of each leg to orient the forces about the crank to produce positive work. Work asymmetry greater than force asymmetry would suggest that, besides a strength/inertial imbalance, one leg is much better at controlling the force direction to effectively produce work (Sanderson, 1990). The force asymmetry of CTA is more than twice that of intact cyclists and may be explained by documented strength imbalance between limbs for persons with amputation (Croisier et al., 2001; Isakov et al., 1996; Pedrinelli et al., 2002). Loss of the ankle joint and the muscles that control it could account, in part, for the strength imbalance. Atrophy of proximal muscles has been documented (Schmalz et al., 2001) and indirectly measured in this study as a difference in thigh circumferences. Considering the thigh of the amputated side is smaller, and that the prosthesis has less mass than the limb it replaced, one should assume there are differences in the inertial properties of each limb. These results, combined with results from other studies (Croisier et al., 2001; Isakov et al., 1996; Pedrinelli et al., 2002; Schmalz et al., 2001), may explain most of the differences in force asymmetry given the limitations of these data. Work asymmetry requires the force to be generated and directed properly to create crank torque. Work production will certainly be influenced by strength and/or inertial differences between limbs but these factors are minor compared with the output of the human system to properly direct the forces for propulsion. Differences in the CTA group were shown between work and force asymmetry, whereas in the intact group both values were similar (Figure 3 and Table 2). The two- to threefold increases in work asymmetry compared with force asymmetry suggests the CTA group had difficulty in the amputated limb not only in generating but more importantly in directing forces at the pedal. These results further suggest a possible change in control strategy allowing the sound limb to compensate for deficiencies in the amputated limb. Greater asymmetry might be considered Figure 5 Foot deflection of the FLEX foot during the TT condition normalized to peak deflection within subject and then averaged across subjects. Shaded area indicates ± 1 SD. Negative numbers indicate FLEX foot compression. The STIFF foot did not show any measurable foot deflection (not shown).

7 320 Childers, Kistenberg, and Gregor a poor result one that is less optimal than desired or it may possibly be optimal given that each individual person has structural and neuromuscular asymmetries at the start (Winter & Sienko, 1988). In other words, the changes in strategy may be making optimal use of existing structural and neuromuscular differences between the amputated limb and the intact limb. Additional research is warranted to determine if the neuromuscular system of a CTA is using the mechanical constraints imposed by the prostheses as well as their remaining neuromuscular and skeletal systems to optimize performance. There was increased variability in work asymmetry observed during the LD condition (Table 2). Sanderson (1991) reported the interaction effect of power and cadence on asymmetry in intact cyclists. Increasing resistance to pedaling requires more of the force to be oriented perpendicular to the crank thereby decreasing the flexibility allowed to the bicycle/rider system to complete the task. Thus, higher difficulty will decrease variability in force production at the pedal as shown in these data. The load conditions were based on a percentage of the subject s predicted maximum heart rate. This meant few subjects pedaled at a similar torque. Although this allowed for a better representation of a nonlaboratory environment, it increased variability in group mean torque production (Figure 4). This increased variability precludes definitive conclusions on pedaling technique of these cyclists. Understanding pedaling technique or how these cyclists compensated within the pedal stroke to use the prosthesis of either foot stiffness would provide great insight into how the CTA group was able to overcome the challenges of the prosthesis. Future research in this area may also expose deficiencies that could be improved with specialized training and/or prosthetic design. Work asymmetry increased as prosthetic foot stiffness decreased during the TT condition related to compression of the prosthetic foot (absorbing energy) during the first 90 of crank rotation and thereby removing energy that would have been transferred into the pedals. The prosthetic foot then decompressed and released that stored energy later in the pedaling cycle ( ) when (1) pedaling forces were relatively less effective at turning the crank and (2) the amount of returned energy was 85% of the energy absorbed. Because energy was removed from the human/bicycle system, the sound limb provided the additional energy to compensate, yielding an increase in work asymmetry during the TT condition. However, pedaling forces were lower during the LD condition and thus not sufficient to deflect the FLEX foot enough to create a difference in work asymmetry. Prosthetic foot stiffness only becomes a factor during high power outputs associated with the simulated time trial condition but not at lower outputs associated with recreational riding. Therefore, if a CTA wishes to cycle in a recreational setting, the prosthetic foot stiffness has little effect and the stiffness properties of a prosthesis designed for walking may be suitable. If a CTA will be cycling at higher power outputs associated with competitive cycling, then the prosthetic foot needs to have greater stiffness. In conclusion, cyclists with a transtibial amputation have greater pedaling asymmetry than intact cyclists and depend more on their sound limb for force and work production. Strength and/or inertial differences between limbs, however, may not fully explain the work asymmetries, thereby suggesting a change in control strategy related to the challenges imposed on the system due to the amputation. Prosthetic foot stiffness increases the work asymmetry in CTA because additional energy is required to compress the foot. The sound limb compensated by producing more torque, thereby increasing work asymmetry for TT level power outputs but not at lower recreational-level power outputs. Future research may address how a CTA is utilizing his or her remaining biological and prosthetic systems to coordinate muscle activation for the generation and distribution of mechanical energy across joints and into the environment for propulsion. Acknowledgments The authors gratefully acknowledge Boris Prilutsky for his review of the manuscript as well as the incredible support from the students, faculty, and staff of the School of Applied Physiology and the subjects who participated in these research studies. We would also like to thank Ossur Corp., Prosthetic Design Inc., Outback Bikes, and Serotta Bicycles for their donation of prosthetic components and bicycle equipment. References Broker, J.P., & Gregor, R.J. (1990). A dual piezoelectric element force pedal for kinetic analysis of cycling. International Journal of Sport Biomechanics, 6, Broker, J.P., & Gregor, R.J. (1996). Cycling Biomechanics. In E.R. Burke (Ed.), High-Tech Cycling (1st ed., pp ). Champaign, IL: Human Kinetics. Carpes, F.P., Rossato, M., Faria, I.E., & Bolli Mota, C. (2007). Bilateral pedaling asymmetry during a simulated 40-km cycling time-trial. Journal of Sports Medicine, and Physical Fitness, 47, Croisier, J.L., Maertens de Noordhout, B., Maquet, D., Camus, G., Hac, S., Feron, F., et al. (2001). Isokinetic evaluation of hip strength muscle groups in unilateral lower limb s. Isokinetics and Exercise Science, 9, Daly, D.J., & Cavanagh, P.R. (1976). Asymmetry in bicycle ergometer pedalling. Medicine and Science in Sports, 8, Faul, F., Erdfelder, E., Lang, A-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, Fox, S.M., III, Naughton, J.P., & Haskell, W.L. (1971). Physical activity and the prevention of coronary heart disease. Annals of Clinical Research, 3, Gailey, R.S. (2004). Physical therapy for sports and recreation. In D.G. Smith, J.W. Michael, & J.H. Bowker (Eds.), Atlas of Amputations and Limb Deficiencies Surgical, Prosthetic, and Rehabilitation Principles (3rd ed., pp ). Rossemont, IL: American Academy of Orthopaedic Surgeons.

8 Cyclists With Unilateral Transtibial Amputation 321 Hafner, B.J., Sanders, J.E., Czerniecki, J.M., & Fergason, J. (2002). Transtibial energy-storage-and-return prosthetic devices: A review of energy concepts and a proposed nomenclature. Journal of Rehabilitation Research and Development, 39, Hunt, M.A., Sanderson, D.J., Moffet, H., & Inglis, J.T. (2004). Interlimb Asymmetry in Persons With and Without an Anterior Cruciate Ligament Deficiency During Stationary Cycling. Archives of Physical Medicine and Rehabilitation, 85, Isakov, E., Burger, H., Gregoric, M., & Marincek, C. (1996). Isokinetic and isometric strength of the thigh muscles in below-knee s. Clinical Biomechanics (Bristol, Avon), 11, Kautz, S.A., & Neptune, R.R. (2002). Biomechanical determinants of pedaling energetics: Internal and external work are not independent. Exercise and Sport Sciences Reviews, 30, Pedrinelli, A., Saito, M., Coelho, R.F., Fontes, R.B.V., & Gaurniero, R. (2002). Comparative study of the strength of the flexor and extensor muscles of the knee through isokinetic evaluation in normal subjects and patients subjected to trans-tibial amputation. Prosthetics and Orthotics International, 26, Perell, K.L., Gregor, R.J., & Scremin, A.M.E. (1998). Lower limb cycling mechanics in subjects with unilateral cerebrovascular accidents. Journal of Applied Biomechanics, 14, Pierson-Carey, C.D., Brown, D.A., & Dairaghi, C.A. (1997). Changes in resultant pedal reaction forces due to ankle immobilization during pedaling. Journal of Applied Biomechanics, 13, Ryan, M.M., & Gregor, R.J. (1992). EMG profiles of lower extremity muscles during cycling at constant workload and cadence. Journal of Electromyography and Kinesiology, 2, Sadeghi, H., Allard, P., Prince, F., & Labelle, H. (2000). Symmetry and limb dominance in able-bodied gait: a review. Gait & Posture, 12, Sanderson, D.J. (1990). The influence of cadence and power output on asymmetry of force application during steadyrate cycling. Journal of Human Movement Studies, 19, 1 9. Sanderson, D.J. (1991). The influence of cadence and power output on the biomechanics of force application during steady state cycling in competitive and recreational cyclists. Journal of Sports Sciences, 9, Schmalz, T., Blumentritt, S., & Reimers, C.D. (2001). Selective thigh atrophy in trans-tibial amputees: an ultrasonographic study. Archives of Orthopaedic and Trauma Surgery, 121, Silverman, A.K., Fey, N.P., Portillo, A., Walden, J.G., Bosker, G., & Neptune, R.R. (2008). Compensatory mechanisms in below-knee gait in response to increasing steady-state walking speeds. Gait & Posture, 28, Smak, W., Neptune, R.R., & Hull, M.L. (1999). The influence of pedaling rate on bilateral asymmetry in cycling. Journal of Biomechanics, 32, Winter, D.A., & Sienko, S.E. (1988). Biomechanics of belowknee gait. Journal of Biomechanics, 21, Wheeler, J.B., Gregor, R.J., & Broker, J.P. (1992). A dual piezoelectric bicycle pedal with multiple shoe/pedal interface compatibility. International Journal of Sports Biomechanics, 8,