Gait and Balance: Gait over Different Terrains (4.3.2.a-e)

For persons with a lower leg amputation (PwA), walking on uneven ground require compensatory movement strategies, partly caused by limited adaptability of prosthetic components. Walking on cross-slopes, a form of uneven ground, is a situation active PwA face regularly. However, only few studies investigated cross-slope walking of PwA [1-3], and the one that studied adaptations focused on the swing phase on the prosthetic side [3].

To investigate in more detail the effects of cross-slopes on gait characteristics (in stance and swing phase) of persons with a transtibial amputation (TTA) and compare it with level walking.

12 TTA were provided with energy-storage and return feet (ESR) for the time of the study. Level walking and cross-slope walking on 5° and 10° tilted tracks were investigated in a gait lab. Kinematic data were recorded with a Vicon system. A Kistler force plate was embedded in the track to measure the ground reaction forces (GRF). Spatio-temporal and biomechanical parameters (GRF, sagittal ankle and knee angles, pelvis inclination and upper body tilt) were analyzed with respect to situation and side. A repeated measurement ANOVA with post-hoc tests were used to show possible differences to level walking.

Spatio-temporal gait characteristics and GRF showed known differences between prosthetic and sound side but reveled only minor adaptations for the different cross-slope situations. Significant adaptations to the different situations were observed in the sagittal knee and ankle angles, in the pelvis inclination and upper body tilt compared to the level situation. The sagittal ankle angle of the sound side showed a reduced range of motion for Valleyside and an increased one for Hillside while the prosthetic side was unaffected. For the sagittal knee angle, adaptations were observed for the sound and the prosthetic side, figure 1A, which adapt the functional leg length. Pelvis inclination and upper body tilt showed similar adaptations to the cross-slope tilt, figure 1B, which are likely compensation and/or stabilization mechanisms of the upper body.

Figure 1: Sagittal knee angle and upper body tilt.

Adaptations to the cross-slopes were observed for the prosthetic and sound side in some parameters. Its effect size increased with increasing cross-slope tilt. The results confirm and extend findings of previous studies [3]. However, most observed effects had the same order of magnitude or were smaller compared to general adaptation effects in gait of TTA if compared to controls.

[1] I. Hak, et al., Arch. Phys. Med. Rehabil. 94 (2013) 2186-93.

[2] I. Starholm, et al., Prosthet. Orthot. Int.; 34(2010) 184-94.

[3] C. Villa, et al., Arch. Phys. Med. Rehabil. 98 (2017) 1149–1157.

Somatosensory feedback is compromised following a limb amputation.  Due to this deficit, lower-limb amputees rely heavily on compensatory mechanisms, such as increased reliance on visual feedback, to maintain balance¹. Our team has developed a sensory neuroprosthesis (SNP) which provides direct sensory feedback to the user based on how the prosthesis interacts with the environment. In this work, we present effects of prolonged home use of the SNP on stair negotiation.

Determine if prolonged use of a SNP at home and in the community could improve performance and alter the compensatory strategy typically utilized by a below-knee amputee to negotiate stairs.

A transtibial amputee with implanted stimulating electrodes used a SNP which provided sensation corresponding to the pressure applied to the plantar surface of the prosthetic foot elicited by electrically stimulating the peripheral nerves in the residual limb².  The participant used the device at home for 291 days and visited the laboratory six times to perform a stair assessment.  The task involved negotiating the 4” and 8” stairs while carrying or not carrying a tray to limit visual feedback.  SNP activated and deactivated conditions were tested with performance order randomized.  Error frequency and strategy were then identified through video analysis.

Three stair negotiation strategies were identified: lead foot preference, foot orientation on step, and stomping. We also identified 10 categories of errors while performing the task which allowed us to quantify any changes in the performance.  Stair height, direction of movement (ascent or descent), SNP mode (activated or deactivated), and tray use influenced strategy selection, error frequency, and type of errors committed.  Although tread depth was uniform, more errors occurred during 4” stair ascent than any other condition.  Each trial was timed, but timing could not differentiate between strategies selected or SNP mode to determine how well the task was performed.  With prolonged home use of the SNP, the subject began to adjust his self-selected compensatory strategies to a more normalized pattern.

Lack of somatosensation causes lower-limb amputees to employ compensatory mechanisms to safely negotiate stairs.  To study how a SNP can influence this task, we systematically characterized stair performance by identifying strategy and error, monitoring changes over time, and quantifying safety through correlation with error frequency.   Our findings suggest that prolonged use of the SNP led to adoption of a more normalized biomechanical approach to stairs and reduction in error frequency.

1. Barnett, C.T., Vanicek, N., Polman R.C.J. Postural responses during volitional and perturbed dynamic balance tasks in new lower limb amputees: A longitudinal study. Gait Posture 37,319–325(2013)
2. Charkhkar, H. et al. High-density peripheral nerve cuffs restore natural sensation to individuals with lower-limb amputations. J. Neural Eng.15,056002(2018)

Supported by Defense Advanced Research Projects Agency and SSC Pacific, Contract No. N66001-15-C-4038 and Department of Defense, Award No. W81XWH-18-1-0321.

Being able to negotiate environmental hazards, such as obstacles on the floor, is important for avoiding the occurrence of a trip or fall and subsequent injury. Previous research has reported on how individuals with both transtibial and transfemoral amputations cross obstacles [1-3]. Information regarding obstacle crossing strategy, lead limb preferences, lower limb joint kinematics and kinetics have been highlighted [1-3]. However, it is not clear how different prosthetic components may influence obstacle crossing behaviour.

To investigate how varying both ankle-foot and knee componentry, in individuals with unilateral transfemoral amputation, influenced their ability to cross an environmental obstacle.

Eight participants (54.8±12.4; 1.8±0.05m; 84.8±13kg; prosthetic experience 24.5±15.91 years) with unilateral transfemoral amputation, crossed an obstacle (0.04x0.08 m; depth*height) placed along a flat walkway, freely choosing their obstacle crossing strategy. Participants used four different combinations of prosthetic componentry; either a microprocessor-controlled (MPK, Orion3) or non-microprocessor-controlled (NMPK) knee component with either a hydraulically articulating (HYD, Echelon) or rigid (RIG, Esprit) ankle-foot component (all Blatchford, Basingstoke, UK). Reflective markers were placed on the lower limbs, with a 13-camera motion capture system (Qualisys AB, Gothenburg, SE) capturing kinematics at 100Hz. Lead limb preference, approach velocity and lead limb vertical toe clearance and final foot placement where recorded.

Self-selected lead limb preferences were unaffected by prosthetic component combination. Only 2/8 participants deviated once (1/4 conditions) from their ‘usual’ lead limb. Approach velocity was highest with the MPK+HYD combination (F(3,18) = 4.16, p = 0.021, ηp² = 0.41). Final foot placement seemed unaffected by prosthetic combination. Although not analysed statistically, vertical toe clearance was lower and varied more by prosthetic combination when leading with the intact vs. prosthetic limb.

Figure 1. Combined limb approach velocity for all combinations of prosthetic componentry.

Regardless of prosthetic component combination, all participants were able to cross the obstacle safely, avoiding trips and falls. In addition, a variety of lead limb preference strategies were observed. This may reflect the relatively high level of function (all K3) and neuromotor flexibility of participants in the study sample. Approach velocity was higher in the MPK+HYD combination, potentially suggesting that obstacle crossing with more functionally advanced prosthetic components could be advantageous, when constraints are placed on task completion time.

1.    Vrieling AH et al. Gait Posture. 2007;26(4):587–94.

2.    Buckley JG et al. J Neuroeng Rehabil. 2013 ;10 ;98.

3.    Barnett CT et al. Pros Orth Int. 2014 ; 38(6) :437-46.

Performing everyday tasks requires accurate foot placement, adequate vision and proprioception [1]. Simulating visual impairment reduces the ability to accurately perceive the characteristics of a raised surface placed within the travel path [2]. Unilateral transtibial limb loss leads to less accurate and precise intact limb foot placement, potentially due to altered proprioceptive feedback [3]. Exploring how impaired vision combined with reduced proprioception, affects obstacle crossing, may contribute to understanding balance and fall-related problems reported in people with limb loss.

To explore the effects of simulated visual impairment on obstacle crossing behaviour in individuals with unilateral transtibial amputation (IUTA).

Seven male IUTAs (38±13 years, 1.83m, 92.5kg, all K3) walked along a 10m walkway crossing an obstacle (two conditions: FLAT = 10mm(H) x 600mm(W) x 100mm(D), TALL = 100mm(H) x 600mm(W) x 10mm(D)) with (blurred glasses, BLUR) and without (clear glasses, CLEAR) simulated visual impairment. Visual acuity and contrast sensitivity were measured in each condition. Full-body kinematics were recorded as participants approached and crossed the obstacle, leading with either the INTACT or PROSTHETIC limb. A three-way repeated measure ANOVA compared differences between limb, obstacle and visual impairment for: lead limb vertical toe clearance (VTC), final foot placement (FFP), and approach velocity.

Visual acuity (Z=2.2, p=0.028) and contrast sensitivity (t(6)=6.24, p<0.001) reduced in the BLUR vs. CLEAR conditions. PROTHETIC limb VTC increased over the FLAT obstacle (+6mm) but reduced over the TALL obstacle (-13mm) compared to the intact limb (F(1,6)=7.57, p=0.03, ηp²=0.56). Lead limb VTC increased for BLUR (115mm) vs. CLEAR (95mm), with this difference being greater in the TALL (+27mm) vs. FLAT (+12mm) obstacle conditions (F(1,6)=42.6, p<0.001, ηp²=0.88). Final foot placement was further away from the obstacle in both the BLUR (207mm vs 165mm) (F(1,6)=19.71, p=0.004, ηp²=0.77) and TALL (202mm vs. 170mm) (F(1,6)=9.732, p=0.021, ηp²=0.62) conditions. Approach velocity reduced when crossing the TALL (1.2m.s-1) vs. FLAT (1.1m.s-1) obstacle (F(1,5)=53.62, p<0.001, ηp²=0.92).

Reduced approach velocity and increased final foot placement when crossing the TALL obstacle suggested that IUTAs perceived the heightened risk of this environmental obstacle. However, VTC was reduced in the PROSTHETIC limb when crossing a TALL obstacle, pointing to an inability to mitigate this risk, potentially increasing the risk of tripping. The increased VTC in the BLUR compared to CLEAR condition over the TALL obstacle also suggests participants acted more cautiously, which may be due to increased task complexity/risk.

1.    Bruijn, S M et al. J R Soc Int 2018; 15; 20170816.

2.    Vale A et al. Ophthal Physl Opt 2008; 28; 135-142.

3.    Foster, R J et al. J Biomech 2020; 105; 109785.

Almost half of persons with major upper limb loss (ULL) fall at least once per year and nearly one third of reported most recent falls result in injury (1). In persons with ULL, the likelihood of falling two or more times in a year increases sixfold with the reported use of an upper limb prosthesis (ULP) (1). These findings, and additional biomechanical studies (2,3), suggest that persons with ULL may experience impaired postural control and locomotor stability.

To characterize the locomotor response of persons with ULL to a simulated trip and assess effects of prosthesis use on that response.

A participant completed two tasks walking on a treadmill (Motek, the Netherlands) wearing a safety harness: 1) baseline walking at 1.0 m/s and self-selected speed, and 2) 12 perturbation trials. Tasks were completed twice: with and without their prosthesis. Perturbation trials consisted of steady-state walking at 1.0 m/s that was unexpectedly interrupted by a rapid treadmill acceleration-deceleration during single limb support on either the sound or impaired side (6 each side). Kinematics were collected with an optical motion capture system (Motion Analysis, CA). Sagittal-plane whole-body angular momentum (WAM), trunk inclination (TI), and trunk inclination velocity (TV) pre- and post-perturbation were estimated using a biomechanical model (Visual3D, C-Motion, MD).

One male participant (58 yrs, 178.0 cm, 95.8 kg, myoelectric prosthesis user) recovered from all perturbation trials without a fall. Instantaneous normalized WAM for one representative perturbation trial is shown in Fig. 1A. Average values for maximum TI, maximum TV and WAM range pre- and post-perturbation across five trials (removing the first from both limb sides) are shown in Fig. 1B. The perturbation successfully generated a forward trunk as reflected by increased TI, TV and WAM. While TI and TV were near equivalent for all conditions, WAM range was considerably higher when the participant did not use their prosthesis to suggest greater postural control demands, and this elevation was greatest when the perturbation was delivered during the sound side single support indicating an asymmetric response.

The increase in TI, TV and WAM post-perturbation indicates that the participant experienced a disturbance to walking behavior from the simulated trip. Importantly, WAM range was greatest post-perturbation when disturbed during sound side single limb support and the participant did not wear his prosthesis. These findings suggest an asymmetric locomotor response following a trip that may share an interaction with ULP wear and could help better understand mechanisms underlying fall risk in this patient group. Data collection is ongoing.

1. Major MJ. Phys Ther 4, 377-387, 2019. 2. Major MJ, et al. Am J Phys Med Rehabil 99, 366-371, 2020. 3. Major MJ, et al. J Electromyogr Kinesiol 48, 145-151, 2019.

This work was supported by the US Department of Veterans Affairs (RX003290).