View Options - Click to expand
Print Options - Click to expand
E-Mail Options - Click to expand

Sense of Feel for Lower-Limb Amputees: A Phase-One Study

John A. Sabolich, CPO
Giovani M. Ortega

ABSTRACT

To determine the effects of a sensory feedback device developed at Sabolich Prosthetic & Research Center for lower-limb amputees, 12 transfemoral (above-knee) and 12 transtibial (below-knee) unilateral amputees were recruited from a convenience population for testing. Pre- and post -testing procedures included. symmetry of weight distribution, duration of single limb standing balance over the involved side, and symmetry of step length and of stance phase times.

After subjects completed the pre-testing protocol, the Sense-of-Feel (SOF) device was incorporated into the pre-test socket and five- to six-hour familiarization periods were provided. Post-testing was performed while wearing the SOF.

Weight distribution symmetry scores increased by 7 percent (p<0.01) overall while the transtibial (below-knee) group improved by 11 percent (p<O.O1). The transfemoral (above-knee) subgroup demonstrated a 24 percent increase (p<O.O4) in standing balance duration. Scores for symmetry of stance phase time increased significantly (p<0.04) for the transfemoral (above-knee) subgroup. For step-length symmetry scores, overall subjects increased by 11 percent (p<0.05), and the transfemoral (above-knee) subgroup improved by 29 percent (p<0.04). Despite the relatively short acclimation period, results suggest that significant change toward improved symmetry was effected by using the SOF. A future study, including a larger sample population using the device for extended periods, is planned to fully understand the benefits of this type of sensory feedback.

Introduction

In recent years, the field of prosthetics has undergone tremendous changes in providing lightweight prostheses with flexible sockets, improved mechanical function and more life-like appearance. The use of advanced materials and improved designs have enabled amputeess to regain more efficient mobility by using these state-of-the art prostheses. These technological advances in materials and design, however, might eventually reach the limit of their capacity to emulate the function of the natural limb.

Humaan locomotion is a dynamic integration of musculoskeletal structure and neurophysiologic function (1). Sensory information from visual, yestibular and somatosensory receptors (including proprioception of the lower limbs and other body parts) is used to select the appropriate motor responses to achieve a smooth, efficient gait. The next generation of lower-limb prostheses should provide the user not only with appropriate biomechanical performance, but also with supplemental sensory information that will enhance functional use of the prosthetic limb.

To date, the role of sensory feedback in the lower-limb prosthesis has received very limited attention, especially when compared to the number of feedback studies for the upper-limb prosthesis (2-6). A full exploration of the benefits of sensory feedback for lower-limb amputees is necessary to definitively establish the need to incorporate this technology into the treatment of amputees.

The loss of motor capability and sensory feedback from the lower limb affects the amputee's ability to establish and maintain his or her center of gravity. For amputees, this deficit is demonstrated in characteristic gait deviations. Clinically, these deviations include asymmetric gait (7-11) and lateral weight-shifting difficulties that result in a halting gait pattern. These asymmetries contribute to a decreased forward velocity and a slower cadence (12-14). Functionally, these deviations translate into increased energy cost of ambulation (15), decreased sense of confidence in using the prosthesis and reduced overall mobility. Medhat (16) concluded that for amputees, physical stamina, balance and gross motor movement presented the greatest difficulties in activities of daily living.

Use of sensory feedback has proven an effective supplement to conventional hands-on therapy in the rehabilitation of balance disorders (17-18). Gapsis, et al., (19) reported that subjects using sensory feedback for controlled weightbearing reached therapy goals in half the time than did those in the control group. Including sensory feedback training during the initial rehabilitation process of a recent amputee could greatly influence the extent of functional outcome in several ways. Sensory feedback training:

  • decreases dependency on visual information
  • helps patients attain therapy goals more quickly (specifically control of lateral weight shifting)
  • improves overall functional independence as a result of increased efficacy of rehabilitation. To date, no known studies have applied sensory feedback to lower-limb amputees during the initial rehabilitation period.

Only a limited number of sensory feedback studies for the lower-limb prosthesis have been reported. Stimulation techniques explored include audio, mechanical and electrical (20). Clippinger, et al., (21) conducted the latest investigation of afferent sensory feedback for the lower-extremity prosthesis. The feasibility of a surgically implanted sciatic nerve stimulator that corresponded to changes in loading patterns on the prosthetic limb was demonstrated. However, for this and similar feedback studies, assessments of feedback device efficacy were obtained mostly from anecdotal reports from the subjects themselves. Other shortcomings of these previous studies were limited sample sizes, non-uniform treatment and no statistical analysis to validate outcomes.

The SOF device is a non-invasive sensory feedback system that provides transcutaneous electrical neural stimulation to afferent sense organs located at the residual limb/socket interface (see Figure 1 ). The sensors, adhered to the plantar surface of the prosthetic foot, send signals proportional in strength to the amount of pressure applied to corresponding electrodes on the residual limb. This configuration enables the user to discriminate between anterior and posterior load conditions along the bottom of the prosthetic foot during the stance phase of the gait cycle.

The hypothesis of this phase-one study was that using the SOF device would effect significant change in gait quality and efficiency as measured by weight distribution, standing balance, and symmetry of stance phase duration and step length. Since amputees generally demonstrate reduced performance in these measures, an improvement in performance may suggest progress toward achieving gait and stance efficiency, which in turn would reduce the possibility of secondary musculoskeletal pathologies resulting from these gait deviations. Perhaps, by collecting external measures representative of internal processing functions, it is possible to gain further insight about the efficacy of the proposed artificial sensory feedback system.

Methods

Subjects were recruited from among patients with scheduled appointments at the Sabolich Prosthetic & Research Center who met inclusion criteria and were willing to volunteer for testing. Candidates using any medication that could affect balance or those requiring use of an assistive device (such as a cane or walker) were not eligible subjects. Candidates were examined by a neurologist to test for vestibular and sensory motor dysfunction and were given a nonfocal exam for normal cerebellum function. Candidates were further examined by an ABC certified prosthetist to ensure an acceptable socket fit to the residual limb and that no skin breakdown or other irritation caused discomfort during stance or gait.

Protocol

Each of the 24 subjects was instructed for all tests. Baseline data were collected for parameter measures as follows:

Weight distribution. Subjects were instructed to place each foot on calibrated, strain gauge-based weight scales in a comfortable position with arms resting at their sides with their eyes fixed on a target on the wall. Continuous readings of vertical force for both sides were recorded in units of pounds over a one-minute period and were calculated as a percentage of body weight.

Single-limb stance. Subjects were instructed to balance as long as possible on their prosthetic limbs. The duration of balance time was measured from the time the subject achieved independent balance to the time the non-supporting limb touched ground. b Ten time trials were measured to calculate mean duration.

ยท Gait analysis. A two-dimensional video gait analysis system was used to collect kinematic gait parameters at a sample rate of 30 frames per second.c Passive reflective markers were adhered directly to the subject's skin over anatomic landmarks commonly used in two-dimensional gait analysis. The subject was then instructed to walk (wearing his or her own shoes) along a marked 50-foot walkway at a natural, casual cadence. This walkway length ensured that transient gait initiation/termination characteristics would not be present during capture of a complete gait cycle. A gait cycle was defined as initiating at heelstrike and terminating at the next heelstrike of the same side.

Five full gait cycles of the involved side and five of the non-involved side were captured. Temporal and spatial information from the digitized data were used to calculate duration of stance time and step length of each side. Stance time, for this application, initiated at heelstrike and terminated at toe-off of the same side and was expressed as a percentage of the gait cycle. Step length was measured as the horizontal distance covered along the plane of progression during double-limb support phase from the toe marker of the rear foot to the heel marker of the front foot. Five step length measures for the involved side and five for the non-involved side were recorded to determine each subject's mean step length.

Subjects were fitted with the SOF device using the same pre-test socket and componentry. Two separate pressure transducers were adhered directly to the bottom of the prosthetic foot. The heel sensor was located 1 cm from the back of the foot, and the "toe" sensor was adhered to what would approximate the apex of the second metatarsal head. Electrodes were then placed on the anterior and posterior aspects of the residual limb so that the signals from the toe and heel sensors were relayed correspondingly. Subjects were instructed to adjust the signal strength for each of the two channels so levels could be perceived as equivalent.

Subjects then familiarized themselves with the functions and characteristics of the SOF device for five to six hours. The Institutional Review Board required the subjects to stay at the center throughout the experiment. The subjects were encouraged to ambulate at 10-minute intervals and to take up to 20-minute rest periods. At the end of this acclimation period, subjects were given the same instructions and performed the same tests as the pre-tests while wearing the SOF device. These measures comprise the post-test measures.

Scoring

Weight distribution scores were calculated as follows:



The weight distribution score (WDS) measures the symmetry between sides. The side bearing less weight (%BW1) is divided by the side bearing the majority of the weight %BW2), resulting in a ratio between 0 and 100 percent. A score of WDS 100 percent would indicate perfect symmetry of weight distribution.

Scores for balance time were based on the mean duration subjects were able to balance on their prosthesis side. Stance time duration for either the involved side (IS) or the non-involved side (NIS) was calculated as a percentage of a complete gait cycle. Five complete gait cycles were captured for each side during pre- and post-testing. Each of the gait cycles was normalized to 100 percent by means of ensemble averaging (22). Mean stance times for pre- and post-testing were calculated and used for scoring. Stance time symmetry scores (STS) were calculated as follows:



The stance time of the IS is divided by the stance time of the NIS. This method for scoring symmetry was selected based on the observation that the stance time of the IS was less than that of the NIS for every subject. A score of 100 would indicate symmetry while a score under 100 would indicate a shorter stance time of the IS than the NIS. To clarify use of this scoring method, suppose pre-testing results in a stance time of 55 percent for IS and 65 percent for NIS. The STS score indicates that IS stance time is 84.6 percent of NIS stance time. Post-test measures of 59 percent for IS and 64 percent for NIS would result in a STS score of 92.2 percent, demonstrating an 8.9 percent increase of symmetry of stance times.

Step-length symmetry (SLS) was calculated as follows:



The shorter step length (SSL) was divided by the longer step length (LSL) regardless of which side (IS or NIS) took the shorter or longer step length A score of SLS - 100 percent indicates perfect step-length symmetry. Statistical significance of change occurring between pre- and post-test measures was determined by paired t-test method. Significance was obtained when p<0.05.

Results

Subject Data

Twelve transfemoral (above-knee) and 12 transtibial (below-knee) unilateral lower-limb amputees participated in the pilot study. Seven of the female subjects were transfemoral, and seven of the male subjects were transtibial amputees. Ages ranged from 21 to 68 years with a mean of 39.5 years (+/- 13.3). The average experience with a prosthesis was 8.1 years (+/- 9.4) with a range of three months to 31 years. Mean weight was 168 lbs. (+/- 40 lbs.) Ten subjects suffered a traumatic amputation; five lost a limb to cancer; seven for dyvascular reasons; and two due to infection. Fifteen of the subjects had left-side amputations while nine had right-side amputations. Subgroups were divided into transfemoral vs. transtibial and male vs. female.

Clinical Data

Results for weight distribution, stance time symmetry and step-length symmetry are found in Figure 2 .

Weight distribution. The overall mean score for symmetry increased from 77 percent to 86 percent (p<0.01). Transtibial scores increased significantly by 19 percent (p<0.01).

Single-limb stance. The transfemoral group increased balance time scores by 24 percent (p < 0.04). Although single-limb stance time for the transtibial group increased by 34 percent, the t-test did not show significance for this increase.

Stance-time symmetry. Mean stance times of the IS were consistently less than those of the NIS for pre- and posttest measures for all subjects. The transfemoral group increased stance time symmetry by 2 percent (p<0.01).

Step-length symmetry. For this measure, the transfemoral group demonstrated a 16 percent increase in scores from 69.5 percent to 80.3 percent (p<0.05).

The compiled scores for all tests for overall performance and the subgroups are listed in Table 1 . Also included are significance levels for each test and the percentage of increase of scores.

Discussion

Using subjects as their own control for this study controlled the effects of factors such as natural ability, prior experience, fatigue, concentration, anxiety and balancing strategies (23). Despite the limited acclimation period, significant improvement in performances suggests that a closer relationship between proprioception and residual motor output was effected for certain subgroups. It is important to note that the subgroups with lower pre-test scores tended to change more than those subgroups with the higher pre-test scores. However, the conclusion that the SOF was useful only for that group must be viewed with caution.

The pilot study data suggest that transfemoral and transtibial patients should probably be treated as coming from distinct populations for the purposes of treatment as well as research design and statistical analysis. In the study samples from each population, the functional limitations of these two groups were quite different at baseline: a multivariate analysis of variance (MANOVA) across four measures yielded F=7.7, df=1,21, p~0.0114 for the overall test of difference; univariate t-tests indicated significant differences on three of four key measures at baseline. When research groups are not comparable at baseline, there is no entirely satisfactory way to scale the measures to render them equal at the start. As a result, differences in amount of change can be very hard to interpret.

As illustrated in Figure 3 , the areas in which transfemoral and transtibial improvement appeared were significantly different too. As depicted, only transfemoral patients improved significantly (at p=0.012) on the measure of step-length symmetry. (In that area transtibial patients actually got a slightly and non-significantly worse score, perhaps in part a function of differences in baseline functioning.) On the other hand, transtibial patients showed significant improvement in weight distribution (p=0.006) while improvement among transfemoral patients on that measure was not significant. Statistical interaction was significant (p=0.025). A similar analysis of the first two principal components in four key outcome measures yielded similar results, suggesting that the way improvement was manifested differed depending on the level of amputation.

Transtibial amputees demonstrated a 19 percent increase in weight-distribution symmetry compared to the 6 percent increase by the transfemoral group. The transtibial group had a lower baseline score than the transfemoral group and thus more room for improvement. Explanation of this outcome could be related to differences in stance behavior between both groups. It is possible that the knee mechanisms in most transfemoral prostheses require weightbearing for increased stability during static stance. Consequently, the transfemoral group may have placed more weight on their prostheses than did their transtibial counterparts during baseline measures.

For single-limb stance the transfemoral group improved by 24 percent. This improvement may be due to increased awareness of how to control the residual limb. It would be interesting to incorporate medial-lateral transducers in addition to the existing anterior-posterior aspects of the feedback system to determine if balance times increase.

The symmetry scores for stance time as well as step length improved significantly for the transfemoral group. Even though the cause for the well known gait asymmetries for amputees may be related to the built-in limitations of the biomechanical performance of the prosthesis itself, it is possible that the sensation provided by the SOF may have increased awareness of timing issues throughout the gait cycle as well as enhanced the kinesthetic cognizance of the prosthetic foot interaction with the ground. The transtibial group's mean baseline scores for stance time and step-length symmetry were significantly greater when compared to the transfemoral group's, suggesting that transtibial amputees walk with a greater symmetry than transfemoral amputees for these measures.

The configuration of the SOF device may have provided features that facilitated to the subjects' rapid accommodation to this type of sensory feedback. Locating sensors to plantar surface sites on the prosthetic foot allowed for argumented feedback critical in heel-strike and toe-off events. The one-to-one correspondence of the sensor-simulator configuration provided a simple means of quickly determining the position and condition of the foot during stance phase. Unlike audio sensory feedback systems, the signal provided by the SOF device required no interpretation of tone or beep to ascertain information regarding foot placement and pressure. Furthermore, the SOF device permitted self-regulation of signal strength by the subject during the acclimation period to obtain threshold signal level suitable for the individual.

Conclusion

Since the direct observation of cerebral and central nervous system (CNS) processing is not possible, the effect an artificial sensory feedback system has on external motor output provides the only means of determining the efficiency of the sensory feedback system. This phase-one study measured parameter changes suggestive of CNS internal response to the SOF device signals by lower-limb amputees. Improved symmetry in weight distribution, step length and stance time were demonstrated for different subgroups. Pretest differences between transfemoral] and transtibial groups indicate that carefully selected test and control group of significant size could provide valuable insight into the use of the SOF by the different subgroups.

Results of the short-term familiarization are encouraging; however, further investigation of long-term use would be necessary to fully explore the benefits of the SOP device. A future study is planned that will implement a study design that enables measurement of specific parameters known to characterize amputee performance, comparison between large control and experimental groups, tracking of performance improvement through repeated measures, and measurement of functional outcome. In this manner, the true value of supplemental sensory feedback for the lower-limb amputee can be determined.

Acknowledgements

We wish to acknowledge that this research was made possible through funding by a Small Business Innovative Research Grant (SBIR) # 1 R43 HD29647-01 through the National Center for Medical Rehabilitation Research of the National Institutes of Health. We also thank Robert Gailey Jr., MSEd, PT, and Mark Thompson, RPT, for their contributions to this grant.


JOHN A. SABOLICH, CPO, is president and clinical director of the Sabolich Prosthetic & Research Center (SPRC), 4301 N. Classen Blvd., Oklahoma City, OK 73118. Sabolich received his degree in prosthetics and orthotics from New York University. He is a clinical instructor for the University of Oklahoma orthopedic surgery and rehabilitation department and an ad]unct associate professor for the University of Oklahoma health sciences center physical therapy department.

GIOVANI M. ORTEGA is director of research and development at the SPRC facility. He received his degree in mechanical engineering from the University of Oklahoma with a background in biomechanics and gait analysis.

References:

  1. Byl N. Spatial orientation to gravity and implications for balance training. In Orthop Phys Ther Clin of N Amer: Exercise Technologies. 1992; 1:2:207-41.
  2. Patterson PB, Katz JA. Design and evaluation of a sensory feedback system tha provides grasping pressure in a myoelectric hand. J Rehab Res Dev 1992;29:1-8.
  3. Riso RR, Ignagni AR, Keith MW. Cognitive feedback for use with FES upper extremity neuroprostheses. IEEE Trans Biomed Eng 1991;38:29-38.
  4. Van Doren CL, Riso RR, Milchus K Sensory feedback for enhancing upper-extremity neuromuscular prostheses. 1991 5:63-74.
  5. Scott RN, Brittain RH, Caldwell RR Cameron AB, Dunfield VA. Sensory-feed back system compatible with myoelectrically control. Med Biol Eng Comput 1980;18 65-9.
  6. Shannon GF. A myoelectrically con trolled prosthesis with sensory feedback. Med Biol Eng Comput 1979;17:73-80.
  7. Zuniga EN, Leavitt LA, Calvert JC. Canzoneri J, Peterson CR. Gait patterns in above-knee amputees. Arch Phys Med Rehab 1972;3:373-82.
  8. Breaky J. Gait of unilateral below-knee amputees. Orth and Pros 1976;30:17-24.
  9. James U, Oberg K. Prosthetic gait pattern in unilateral above-knee amputees. Scand J Med 1973;5:35-50.
  10. Robinson JL, Smidt GL, Arora JS. Accelerographic, temporal and distance gait factors in below-knee amputees. Phys Ther 1977;57:898-904.
  11. Murray MP, Sepic SB, Gardner GM, Mollinger LA. Gait patterns of above-knee amputees using constant friction knee components. Bull Pros Res 10-34, 1981;17:2:3545.
  12. Waters RL, Perry J, Antonelli D, Hislop H. Energy cost of walking of amputees: The influence of level of amputation, JBJ S 1976;58A:42-6.
  13. Huang CT, Jackson JR, Moore NB, Fine PR, Kuhlemeier KV, Traugh GH, Saunders PT. Amputation: energy cost of ambulation. Arch Phys Med Rehab 1979;60: 18-24.
  14. Tiberwala DN, Ganguli S. Pattern recognition in tachographic gait records of normal and lower-extremity handicapped human subjects. J Biomed Eng 1982;4: 233-40.
  15. Skinner HB, Effeney DJ. Gait analysis in amputees. Am J of Phys Med 1985; 64:82-9.
  16. Medhat A, Huber PM, Medhat MA. Factors that influence the level of activities in persons with lower-extremity amputation. Rehab Nursing 1990;15:1:13-8.
  17. Wannstedt GT, Herman RM. Use of augmented sensory feedback to achieve symmetrical standing. Phys Ther 1978; 58:5:553-9.
  18. Clarke AH, Krzok W, Scherer H. Posturography with sensory feedback-a useful approach to vestibular training? In: Disorders of Posture and Gait. Brandt T. Paulus W, Bles M, Dieterich M, Krafczyk 5, Straub A [eds]. Georg Thieme Verlag Stuttgart. New York, 1990.
  19. Gapsis JJ, Grabois M, Borrell RM, Menken SA, Kelly M. Limb load monitor: evaluation of a sensory feedback device for controlled weightbearing. Arch Phys Med Rehab 1982;63:38-41.
  20. Clippinger FW, McElhaney JH, Maxwel MG, Vaughn DW, Horton G, Bright L. Prosthetic sensory feedback lower extremity. Newsletter: Pros and Orth Clin 1981;5:3:1-4.
  21. Clippinger FW, Seaber AV, McElhaney JH, Harrelson JM, Maxwell MG. Afferent sensory feedback for lower-extremity prosthesis. Clinical Orth and Related Research 1982;169:202-6.
  22. Winter DA. The biomechanics and motor control of human gait. 1991. University of Waterloo Press. 2nd ed. Ontario, Canada,1991:9.
  23. Goldie PA, Evans GM, Bach TM. Steadiness in one-legged stance: development of a reliable force-platform testing procedure. Arch Phys Med Rehab 1992; 73:348-54.