Gait Analysis in Prosthetics by James R. Gage, M.D. Ramona Hicks, R.P.T., M.A.

REVIEW lems faced by lower amputees. Inman's measurement techniques included motion pic­ Objective measurement systems which quan­ tures of coronal and sagittal views, as well as tify locomotion have been in use for the past transverse rotations from below using a glass century. But not until World War II, when walkway. Using interrupted light photography, thousands of men returned home to the United the Laboratory team studied the States with , was technology really motion of body segments during . Force applied to the understanding of prosthetic gait. plates measured the subject's ground reaction Inman and colleagues1 founded the Biome­ forces, and muscle activity was recorded using chanics Laboratory at the University of Cali­ electromyography (EMG), which measures the fornia to establish fundamental principles of electrical signals associated with contraction of human , particularly in relation to prob­ a muscle. Prior to Inman's fundamental studies, prostheses were customized for the individual Temporal and kinematic data, which were col­ amputee, without any particular regard to ra­ lected at slow, free, and fast speeds, showed tional structural design. Inman's goal was to that the hydraulic improved the provide fundamental data essential for the de­ between the prosthetic limb and the sound limb, sign of prosthetic limbs. By analyzing normal especially at the fast and free speeds. This human walking, he and his colleagues laid the finding was true for both cadence and the groundwork for biomechanical analysis of am­ amount of -flexion at swing phase. putee gait.2 Since that time, numerous tech­ Hoy and colleagues,10 in one of the few niques have been developed to study human lo­ studies on gait in juvenile amputees, collected comotion,3 and numerous studies have been un­ kinematic data at various speeds to compare the dertaken to evaluate prosthetic gait. solid cushioned (SACH) to a Child Amputee Prosthetic Project (CAPP) ex­ perimental foot. The authors found range RESEARCH APPLICATIONS of motion to be closer to normal and signifi­ Eberhart, et al.4 described the locomotor cantly less with the CAPP foot than the SACH mechanism of the above-knee amputee from ki­ foot. 11 nematic and kinetic data. They compared lateral Hannah and Morrison studied the effect of stick figures of amputees to normal subjects as alignment of the below-knee on gait. a means to objectively identify Using electrogoniometers to measure hip and in the sagittal plane. Force plate data were used knee joint rotations in the coronal, sagittal, and to compare the weight-bearing characteristics of transverse planes, they found that malalignment the prosthetic limb and the sound limb. From of the prosthetic foot was the most crucial for these comparisons, the authors identified am­ gait symmetry. 12 putees who walked well with their prostheses Grevsten and Stalberg used electromyog­ and those who were less adept. Eberhart be­ raphy to compare muscle activity in below-knee lieved that ultimately "optimal" patterns of amputees walking with patellar tendon-bearing gait could be determined for amputees and used (PTB) and PTB-suction prostheses. Surface as a reference for evaluating prosthetic gait. electrodes were placed over the tibialis anterior Zuniga, et al.5 studied gait in 20 above-knee and gastrocnemius muscles which, in normal amputees by using electrogoniometers attached gait, usually fire at opposite phases. The data to the knee and foot switches. Their data doc­ showed that these muscles contracted for longer umented in the stance and swing periods when the PTB prosthesis was used than phase times between the prosthetic and sound with the PTB-suction prosthesis, suggesting limb. that the suction mechanism improved the adap­ In similar investigations, James and Oberg6 tation to the prosthesis. 13 and Murray, et al.7 studied temporal stride pa­ Thiele, et al. investigated possible neuro- rameters and knee flexion-extension angles, physiological reasons for weakness in above- and also examined above-knee gait at various knee amputees by recording electromyographic speeds. They confirmed the stance and swing activity of the quadriceps during gait. They did phase asymmetry between the prosthetic and not find abnormal recordings and concluded sound limb. They also showed that the asym­ that muscle weakness was secondary to bio­ metry was present regardless of the speed of mechanical, rather than neurophysiological, walking. factors. The collection of baseline data in above-knee amputees clearly demonstrated some shortcom­ CLINICAL APPLICATIONS ings in prosthetic gait. One of these, the longer swing time which is required on the prosthetic Until the present, has been ap­ side, has led to the development of dozens of plied to prosthetics only for research purposes. prosthetic knees. Gait analysis laboratories Routine prosthetic fitting and checkout are still have been used to evaluate some of these pros­ done by means of observational gait analysis. thetic designs. Godfrey, et al.,8 in a limited However, observational gait analysis has many study that compared gait with six cadence-re­ disadvantages. sponsive knee units, found no significant dif­ In the first place, even normal human ferences among them. Murray, et al.9 compared walking is extremely complex. With each step, the gait of above-knee amputees with hydraulic more than 30 major muscles have to contract knee units versus constant friction knee units. and/or relax synchronously in each lower ex- tremity. Also, normal human gait is rapid (ap­ this objective analysis is produced by the com­ proximately 105 steps per minute), and the puter in such a fashion that preconceived biases is not fast enough to separate the and communication errors between observers various components of gait at this speed. Krebs, are minimized. Furthermore, some of the more et al.14 have shown that data vary widely when modern gait analysis facilities have the ability different examiners have observed a person's to compare records, for example, of a patient's gait and that observational analysis is only a gait pre- and post-operatively, or of an ampu­ moderately reliable technique. The variations tee's gait with two different prosthetic devices between observers may be due to the precon­ or components. Through comparisons like ceptions of individual observers, to limitations these, the presence or absence of benefit can be of human perception, or to problems in trans­ determined objectively. mitting the information or data to colleagues. In light of these findings, it is not surprising that the fit and quality of the limbs fabricated by different prosthetists vary greatly. KINESIOLOGY Technology has now progressed to the point The field of prosthetics can make use of the where automated gait laboratories can be built. new science of kinesiology, or the study of Their capabilities vary, but most labs monitor movement. Kinesiology consists of two major one or more of the following parameters: fields: 1) kinematics or movement measurements 1) kinematics, the study of motion exclusive through a motion analysis system, of the influences of mass or forces, i.e., 2) evaluation of ground reaction forces via without regard to the underlying of force plates or pressure sensitive the motion; and switches, and 2) kinetics, which deals with the forces that 3) dynamic electromyography (monitoring produce motion. the electrical activity of contracting mus­ cles). Kinematic Data The advantage of an automated motion mea­ Kinematic data can be gathered in a variety surement system is that automated data entry of ways—through interrupted light photog­ and rapid processing allow routine clinical use raphy, cinefilm, video systems, and/or electro- at a reasonable cost. Since the sampling rate of goniometers—and it can be displayed in many most automated motion systems is in excess of ways. Stick figures provide a visual display of 50 Hz (50 samples/second), all movement in the subject walking. the lower extremities during walking can be ex­ Figure 1 is a stick figure representation of an amined in detail and with excellent reproduc­ 11 year old girl with a right knee disarticula­ ibility. tion. The stick figures facilitate the identifica­ Thus, the analysis of walking becomes ob­ tion of gait deviations, e.g., knee hyperexten- jective, rather than subjective, and a record of sion on the prosthetic side at stance phase. With

Figure 1. Lateral stick figures of the right gait cycle of an 11-year-old-girl with a right knee disarticulation. Figure 2. Comparison of knee flexion-extension motion in one above-knee amputee with an average composite of knee flexion- extension in seven above- knee amputees.

observational gait analysis, this gait deviation Although all above-knee amputees hyperextend might be missed, or two examiners might argue their knees slightly during stance phase, this about its presence. With objective gait analysis, patient has 10 degrees more hyperextension we can prove the deviation's existence by than average. Following the gait analysis, it viewing the stick figures, and we can identify was discovered that the knee extension bumper the cause of the deviation by reviewing the was too soft, and it was replaced with a suff­ graphs that depict motion. These graphs display er one. motions of each joint of the lower extremities Kinematic data can also be used to compute in all three planes during a representative gait temporal data, such as stride length, cadence, cycle. and walking velocity. Figure 3 compares the Figure 2 is a graph showing knee flexion- temporal data of the child with knee disarticu­ extension of the same child with a knee disar­ lation with "normal" children the same age. ticulation. The child's sagittal knee motion is Notice that the stride length is normal but that compared with the mean or average flexion- the walking velocity and cadence are less than extension of seven other above-knee amputees. normal.

Figure 3. Linear measurements of an 11-year-old girl with a knee disarticulation compared with a composite of linear measurements of normal children. Figure 4. Graphic display of the vertical ground reaction forces in a 9-year-old boy with and without shoes.

Kinetic Data foot. Also, notice with the SACH foot how the initial forces move from an anterior to posterior The forces that cause movement are usually direction as the heel compresses. This pattern collected through pressure sensitive switches or is not seen in the multi-axis foot. paper, or with commercial force plates, which are designed to break down the ground reaction forces into their components (X,Y,Z force, and DYNAMIC X,Y,Z moment). The software of a modern gait analysis laboratory is able to combine force ELECTROMYOGRAPHY plate data with motion analysis data to produce Dynamic electromyography is a valuable tool meaningful graphic outputs. for measuring the time duration of muscle ac­ Figure 4 shows the vertical ground reaction tivity, which is recorded through electrodes, ei­ force (Z force) for walking barefoot compared ther surface or indwelling. However, since vol­ with walking with shoes in a 9 year old boy untary muscle activity results in an electromyo­ with a Symes prosthesis. Notice the improved graphic recording that increases in magnitude symmetry at push-off between the prosthetic with the tension, other variables can also influ­ and sound limb when shoes are worn. ence the signal, limiting the accuracy of EMG Force plates can also be used to compute the as a predictor of muscle tension. location of the center of pressure on the foot. Electromyographic data can be displayed in Figure 5 compares the foot force progression several ways. When used to analyze a gait pattern of a SACH foot to a multi-axis foot in cycle, the data show which muscles are active a 27 year old male with a below-knee ampu­ during each phase of gait. Figure 6 compares tation. From these data, one can see that the muscle activity during gait of the subject foot force progression pattern is more lateral walking with the SACH foot compared with the with the multi-axis foot than with the SACH multi-axis foot. The hamstrings and quadriceps

Figure 5. Path of the center of pressure on the foot in a 27-year- old below-knee amputee with a SACH foot and with a multi-axis foot. Figure 6. EMG activity of the ham­ strings and quadriceps muscles during gait in a 27-year-old patient with a SACH foot and with a multi- axis foot.

muscle groups were sampled and show the rated into dynamic alignment of each new pros­ same firing patterns regardless of the type of thesis, helping to insure appropriate alignment foot that is worn. What is interesting is that the and fit. Finally, prosthetic research, using both hamstrings are firing just before -off when kinematics and kinetics, will continue as we they are usually silent and the quadriceps are seek to identify and rectify the problems created inactive at this time when normally they fire to by loss of the body's normal limb. The ultimate restrain knee flexion and prevent excessive heel outcome of this research will be the develop­ rise. As might be expected, this patient walks ment of components that will be stronger, with exaggerated knee flexion at swing phase. lighter in weight, and much more functional than those used now. SUMMARY Gait analysis is useful in evaluating an am­ putee's prosthesis by providing objective mea­ REFERENCES surements and a permanent record of the pa­ 1 Fundamental Studies of Human Locomotion and Other tient's status. Kinematic, kinetic, and EMG Information Relating to Design of Artificial Limbs. Sub­ data assist the clinician and prosthetist in iden­ contractor's Report to the Committee on Artificial Limbs. tifying specific problems encountered by the National Research Council. Prosthetic Devices Research amputee and in identifying the causes. Gait Project, College of Engineering, University of California, Berkeley. Serial No. CAL 5. 2 vols. The Project, Berkeley, analysis also allows comparison of different 1947. prosthetic designs or different alignments of the 2 Radcliffe, C.W., "Functional considerations in the fit­ same prothesis. Most importantly, however, the ting of above-knee prostheses," Selected Articles From Ar­ record provided allows examiners to objectively tificial Limbs, Huntington, NY, Kreiger Publishing Co, Inc, discuss the problems and their potential solu­ 1970, p.p. 5-30. 3 Winter, D.A., Biomechanics of Human Movement, tions. New York, John Wiley & Sons, 1979, p.p. 9-46. 4 Eberhart, H.D.; Elftman, H.; and Inman, V.T., "The locomotor mechanism of the amputee," Klopsteg PE, FUTURE APPLICATIONS Wilson PD, et al (eds): Human Limbs and Their Substitutes, New York, Hafner Publishing Co, 1968, p.p. 472-480. The field of prosthetics will begin to change 5 Zuniga, E.N.; Leavitt, L.A.; Calvert, J.C.; Canzoneri, rapidly with the application of kinesiology. J.; and Peterson, C.R., "Gait patterns in above-knee am­ Soon, optimal standards of gait will be estab­ putees," Arch Phys Med Rehabilitation, 53:373-382, 1972. lished for each prosthetic level. With the wide­ 6 James, U. and Oberg, K., "Prosthetic gait pattern in spread availability of low-cost motion analysis, unilateral above-knee amputees," Scand J Rehabil Med, kinematic analysis will be routinely incorpo­ 5:35-50, 1973. 7 Murrary, M.P.; Sepic, S.B.; Gardner, G.M.; and Mol­ 12 Grevsten, S. and Stalberg, E., "Electromyographic linger, L.A., "Gait patterns of above-knee amputees using study of muscular activity in the stump while constant-friction knee components," Bull Prosthet Res, walking with PTB- and PTB-suction prosthesis," Ups J 17(2):35-45, 1980. Med Sei, 80:103-112, 1975. 8 Godfrey, C.M.; Jousse, A.T.; Brett, R.; and Butler, 13 Thiele, B.; James, U.; and Stalberg, E., "Neuro- J.F., "A comparison of some gait characteristics with six physiological studies on muscle function in the stump of knee joints," Orthotics and Prosthetics, 29(3):33-38, above-knee amputees," Scand J Rehabil Med, 5:67-70, 1975. 1973. 9 Murray, M.P.; Mollinger, L.A.; Sepic, S.B.; Gardner, 14 Krebs, D.E.; Edelstein, J.E.; and Fishman, S., "Re­ G.M.; and Linder, M.T., "Gait patterns in above-knee am­ liability of observational kinematic gait analysis," Ac­ putee patients: Hydraulic swing control vs constant-friction cepted for publication in J Phys Ther, 1985. knee components," Arch Phys Med Rehabil, 64:339-345, 1983. 10 Hoy, M.G.; Whiting, W.C.; and Zernicke, R.F., "Stride kinematics and knee joint kinetics of child amputee gait," Arch Phys Med Rehabilitation, 63:74-82, 1982. AUTHORS 11 Hannah, R.E. and Morrison, J.B., "Prostheses align­ James R. Gage, M.D. and Ramona Hicks, R.P.T., M.A. ment: Effect on gait of persons with below-knee amputa­ are with the Kinesiology Laboratory at Newington Chil­ tions," Arch Phys Med Rehabil 65:159-162, 1984. dren's Hospital in Newington, Connecticut 06111.