Artificial intelligence (AI) algorithms for gait analysis rely on properly identified foot strikes for step-based feature calculation. Smartphone signals collected during movement assessments, such as the 6-minute walk test (6MWT), have been used to train AI models for foot strike identification and fall risk classification in able-bodied populations. However, there is limited research in populations with more asymmetrical gait. People with lower limb amputation can have high gait variability, adversely affecting automatic step detection algorithms. Hence, fall risk models for lower limb amputees have relied on manual foot strike labelling to calculate step-based features for model training, which is inefficient and impractical for clinical use. In this thesis, decision tree and long-short term memory (LSTM) models were developed, optimized, and their performance compared for automated foot strike identification in an amputee population. Eighty people with lower limb amputations (27 fallers, 53 non-fallers) completed a 6MWT with a smartphone at the posterior pelvis. Automated and manually labelled foot strikes from the full 6MWT and from the first two minutes of data were used to calculate step-based features. A random forest model was used to classify fall risk. The best foot strike identification model was an LSTM with 100 hidden nodes in the LSTM layer, 50 hidden nodes in the dense layer, and batch size of 64 (99.0% accuracy, 86.4% sensitivity, 99.4% specificity, 82.7% precision). Automated foot strikes from the full 6MWT data correctly classified more fallers (55.6% versus 48.1%), whereas automated foot strikes from 2-minute data classified more non-fallers (90.6% versus 81.1%). Feature calculation using manually labelled foot strikes resulted in the best overall performance (80.0% accuracy, 55.6% sensitivity, 92.5% specificity). This research created a novel method for automated foot strike identification in lower limb amputees that is equivalent to manual labelling and demonstrated that automated foot strikes can be used to calculate step-based features for fall risk classification. Integration of the foot strike identification model into a smartphone application could allow for immediate stride analysis after completing a 6MWT; however, fall risk classification model improvement is recommended to enhance clinical viability.

This study was conducted to examine the perception of others towards individuals with upper limb amputation and to determine how their ratings were affected by the presence of a prosthesis. The survey included 469 participants from a university in the southeastern United States. Participants read a brief background scenario and then rated pictured individuals with or without amputation, and with and without prostheses. Our assumption was that if one type of image was rated more negatively than others, this discrepancy would quantify stigma and stereotype. After viewing the images, participants rated several attributes associated with disability stereotypes and perceived functional ability using a semantic differential and ability rating scale. Multiple significant findings were observed including higher competence, warmth, and ability ratings of amputees with prostheses as compared to both non-amputees and amputees, lower competence and ability ratings in females as compared to males, and higher ability ratings for individuals with myoelectric prostheses as compared to body-powered prostheses. Perceptions of amputees are important in understanding amputees' experience in many areas such as employment opportunities and psychosocial functioning.

No description available.

Several studies have been devoted to the metabolic costs of amputees walking on prostheses with different masses added to their components. However, limited study has been directed at quantifying the mass differences of the actual materials available to amputees and the metabolic and mechanical work required to walk on these materials. The energetics of two materials currently used in the design of lower extremity prosthetics were examined in an attempt to determine if mass differences had an effect on amputee walking. A total of fifteen, unilateral amputees (8 transfemoral and 7 transtibial) performed treadmill walking on prostheses assembled from titanium and stainless steel components. Standardized components (knees, pylons, adapters, feet) made from each material were added below the level of the socket. Submaximal oxygen consumption {W/kg} and mechanical power allowing transfers within and between segments {W/kg} were calculated as subjects walked at self-selected velocities until steady state was achieved. Results show that despite significant mechanical differences [F(1,12)= 4.85, p<.048], the decreased mass associated with the use of titanium materials does not have an effect on the metabolic costs [F(1,14)=1.45, p<.249] of the subjects in this study. In addition, stride rate and stride length showed little differences when walking with both materials. Further division of subjects by age and experience walking on a prosthesis do suggest that older amputees and established walkers do benefit most from the use of titanium, both metabolically and mechanically. The choice of materials for use in every day walking will display differences in the mechanical work of amputees however, these differences are not great enough to realize metabolic consequences. / Thesis / Master of Science (MS)

A rear locking prosthetic knee joint with a durable, rear Automatic Stance-Phase Lock (ASPL), was developed to investigate the versatility of the (ASPL) mechanism in improving the functionality of prosthetic knees appropriate for a global market. An international survey and a Quality Function Deployment identified deficits with existing prosthetic knee mechanisms and established the most influential design parameters. Work on the knee design was completed following a comparative stability analysis of different knee mechanisms which justified the initial design. Solid models were generated with computer design software and a prototype was produced and structurally tested. Finally, clinical pilot testing was conducted on a unilateral transfemoral amputee, and various gait variables were assessed. As hypothesized, the knee performed close to the level of a conventional six-bar knee providing highly effective stance-phase control and the pilot test showed that improvements to the swing-phase response could further reduce the asymmetry of gait.

A rear locking prosthetic knee joint with a durable, rear Automatic Stance-Phase Lock (ASPL), was developed to investigate the versatility of the (ASPL) mechanism in improving the functionality of prosthetic knees appropriate for a global market. An international survey and a Quality Function Deployment identified deficits with existing prosthetic knee mechanisms and established the most influential design parameters. Work on the knee design was completed following a comparative stability analysis of different knee mechanisms which justified the initial design. Solid models were generated with computer design software and a prototype was produced and structurally tested. Finally, clinical pilot testing was conducted on a unilateral transfemoral amputee, and various gait variables were assessed. As hypothesized, the knee performed close to the level of a conventional six-bar knee providing highly effective stance-phase control and the pilot test showed that improvements to the swing-phase response could further reduce the asymmetry of gait.

Unilateral below-knee amputees commonly experience asymmetrical gait patterns and develop comorbidities in their intact (non-amputated) and residual (amputated) legs, with the mechanisms leading to these asymmetries and comorbidities being poorly understood. Prosthetic feet have been designed in an attempt to minimize walking asymmetries by utilizing elastic energy storage and return (ESAR) to help provide body support, forward propulsion and leg swing initiation. However, identifying the influence of walking speed and prosthetic foot stiffness on amputee gait mechanics is needed to develop evidence-based rationale for prosthetic foot selection and treatment of comorbidities. In this research, experimental and modeling studies were performed to identify the influence of walking speed and prosthetic foot stiffness on amputee walking mechanics. The results showed that when asymptomatic and relatively new amputees walk using clinically prescribed prosthetic feet across a wide range of speeds, loading asymmetries exist between the intact and residual knees. However, knee intersegmental joint force and moment quantities in both legs were not higher compared to non-amputees, suggesting that increased knee loads leading to joint disorders may develop in response to prolonged prosthesis usage or the onset of joint pathology over time. In addition, the results showed that decreasing ESAR foot stiffness can increase prosthesis range of motion, mid-stance energy storage, and late-stance energy return. However, the prosthetic foot contributions to forward propulsion and swing initiation were limited due to muscle compensations needed to provide body support and forward propulsion in the absence of residual leg ankle muscles. A study was also performed that integrated design optimization with forward dynamics simulations of amputee walking to identify the optimal prosthetic foot stiffness that minimized metabolic cost and intact knee joint forces. The optimal stiffness profile stiffened the toe and mid-foot while making the ankle less stiff, which decreased the intact knee joint force during mid-stance while reducing the overall metabolic cost of walking. These studies have provided new insight into the relationships between prosthetic foot stiffness and amputee walking mechanics, which provides biomechanics-based rationale for prosthetic foot prescription that can lead to improved amputee mobility and overall quality of life. / text

Background There are more than 300,000 persons in the U.S. living with transfemoral amputation (TFA). Persons with TFA use a knee prosthesis for gait and mobility. Presently, the C-Leg microprocessor knee prosthesis is the standard of care. C-Leg has significantly improved safety and cost efficacy and has created modest gains in gait efficiency. Recently, a new prosthesis has introduced a new sensor array and processor that reportedly improves knee motion, stair function and standing stability. Early claims of the reported functional benefits of the new Genium knee (formerly X2) have not been validated in a rigorous clinical trial. Therefore, the purpose of this project was to determine if the Genium knee improves safety, function and quality of life compared to the current standard of care (C-Leg). Methods The study is a randomized AB crossover with a control group. Subjects must have used (and still be using) a C-Leg for a minimum of 1yr prior to enrollment. Inclusion criteria beyond this are unilateral transfemoral or knee disarticulation amputation for any etiology, community level ambulation (Medicare level 3 or above), independent ambulation and ability to independently provide written, informed consent. Once enrolled subjects utilize their same socket but receive a study foot (Trias or Axtion). Subjects are randomly assigned to either stay with their C-Leg or be fit with a Genium knee. Subjects accommodate and test (A phase) then crossover to the other knee condition and repeat the testing (B phase). A follow up phase of the study beyond the B phase is ongoing to study longer term preference. For AB assessment, three domains were assessed: Safety, function and quality of life. For safety, the PEQ-A survey of stumbles and falls, posturography (Biodex SD limits of stability and postural stability tests), 4 square step test and 2 minute ramp stand test were completed. For function, a series of timed walking tests, the amputee mobility predictor, kinematic gait assessment and physical functional performance-10 tests were conducted. For quality of life, the socioemotional and situational satisfaction domains of the population specific and validated PEQ (prosthesis evaluation questionnaire) were completed. Results Safety: Posturographic assessment revealed impairment between transfemoral amputees and non-amputees. Stumbles and semi-controlled falls decreased with Genium but were not significantly different. Four square step testing was significantly (p 0.05) improved from 12.2s(3.3) to 11.1s(3.4) for the C-Leg and Genium respectively. Function: Kinematic asymmetry was minimally different between knee conditions. The AMP mean(SD) scores while subjects used C-Leg was 40.8(3.6; 33-45) and 43.3(2.6) [p<0.001]. PFP scores (cumulative), upper body function and endurance scores were improved with Genium compared with C-Leg at 9.1%(p=0.03), 8.7%(0.01) and 10.3%(0.04) respectively. Quality of Life: For quality of life, situational satisfaction favored Genium (p<0.001) which included subject's satisfaction with gait, training and quality of life in general. Conclusion C-Leg and Genium promote static weight bearing beyond asymmetric values reported in the literature. In terms of limits of stability, TFA's are clearly impaired, primarily over the amputated side posteriorly however the Genium seems to enable posterior compensations that coincide with multi-directional stepping improvements. Anteriorly, the C-Leg's toe triggering requirements seem to improve limits of stability but come at the cost of discomfort on ramp ascent. With regard to safety, it seems that both knee systems represent good options for the community ambulating TFA. The largest improvements with Genium were in the activities of daily living assessment; predominantly balance and upper body function. It seems that the combination of multi-direction stepping with starts and stops and stair ascent are key areas of improvement. In conclusion, the sensor array in the Genium knee prosthesis promotes improved function in activities of daily living. Specifically improved in this context were balance, endurance, multi-directional stepping, stair ascent and upper limb function in highly active transfemoral amputees.

Partial foot amputation is becoming a more viable and common surgical intervention for the treatment of advanced diabetes, vascular insufficiency and trauma. Statistics describing the incidence of partial foot amputation are scarce. In Australia, it is not known how many people undergo partial foot amputation annually however in the United States upwards of 10,000 partial foot amputations are performed each year. Many of these procedures are likely to be in preference to below-knee amputation under the pretext of improved function associated with preserving the ankle joint and foot length despite common failings including ulceration and equinus contracture which can lead to more proximal amputation. There is a substantial body of literature, which lends support to the contention that much of clinical practice has not been based on experimental evidence describing the gait of partial foot amputees or the influence of prosthetic and orthotic intervention. This limited scientific underpinning of practice may contribute to the common failures and allow misconceptions, such that preserving foot length and the ankle joint improves function, to perpetuate. The aim of this investigation was to develop accurate mechanical models to analyse the effects of amputation and prosthetic/orthotic intervention on the gait of partial foot amputees. Anthropometric and linked-segment inverse dynamic models were developed to accurately depict the affected lower limb and account for prosthetic/orthotic intervention and footwear. These novel techniques enhance the accuracy of kinetic descriptions, affecting the results obtained for terminal swing phase. These models more accurately portray the requirements of the hamstring and gluteus maximus muscles to decelerate the swinging limb in response to the net increase in mass and inertia of the limb segments due to prosthetic fitting. With an appreciation of the influence these models have on the estimation of kinetic parameters, the gait of partial foot amputees was investigated. Kinematic abnormalities were primarily limited to the ankle and were characterised by poor control of tibial rotation during the mid-stance phase consistent with reduced eccentric work by the triceps surae muscles. The centre of pressure excursion and anterior progression of the trunk outside the reduced base of support was limited until contralateral initial contact; which could reflect triceps surae weakness and an inability to substantially load the prosthetic forefoot. Reductions in power generation across the affected ankle were the result of reductions in the angular excursion of the ankle and reductions in the ankle moment. Reductions in the ankle moment were consistent with the limited excursion of the centre of pressure commensurate with peak ground reaction forces. During early stance, concentric activity of the hip extensor musculature was observed, bilaterally, to advance the body forward. Results from these investigations focus on restoring power generation across the ankle given that the primary reason for preserving the ankle joint and calf musculature would seem to be the ability to use it functionally. Improvements in triceps surae strength may allow individuals to capitalise on improvements in below ankle prosthetic design and affect significant improvements in ankle power generation. In conjunction with improvements in muscle strength, below ankle prosthetic design needs to incorporate a socket and toe lever capable of comfortably distributing forces caused by loading the prosthetic forefoot. In conjunction with improvements in muscle strength, above ankle prosthetic design needs to incorporate an ankle joint. The development of a suitable joint poses significant design challenges for the engineer and prosthetist. This thesis provides new insights into the gait of partial foot amputees and the influence of prosthetic/orthotic design, which challenge common misconceptions underpinning clinical practice, prosthetic prescription and surgery. Aside from advancing the understanding of partial foot amputee gait and the influence of prosthetic/orthotic fitting, these investigations challenge and aim to improve current prosthetic and rehabilitation practice. Thus reducing the incidence of complications, such as ulceration which have been associated with the need for more proximal below knee amputation and allow partial foot amputees to utilise the intact ankle joint complex.

Falls are a leading cause of injury and accidental injury death worldwide. Fall-risk prevention techniques exist but fall-risk identification can be difficult. While clinical assessment tools are the standard for identifying fall risk, wearable-sensors and machine learning could improve outcomes with automated and efficient techniques. Machine learning research has focused on older adults. Since people with lower limb amputations have greater falling and injury risk than the elderly, research is needed to evaluate these approaches with the amputee population. In this thesis, random forest and fully connected feedforward artificial neural network (ANN) machine learning models were developed and optimized for fall-risk identification in amputee populations, using smartphone sensor data (phone at posterior pelvis) from 89 people with various levels of lower-limb amputation who completed a 6-minute walk test (6MWT). The best model was a random forest with 500 trees, using turn data and a feature set selected using correlation-based feature selection (81.3% accuracy, 57.2% sensitivity, 94.9% specificity, 0.59 Matthews correlation coefficient, 0.83 F1 score). After extensive ANN optimization with the best ranked 50 features from an Extra Trees Classifier, the best ANN model achieved 69.7% accuracy, 53.1% sensitivity, 78.9% specificity, 0.33 Matthews correlation coefficient, and 0.62 F1 score. Features from a single smartphone during a 6MWT can be used with random forest machine learning for fall-risk classification in lower limb amputees. Model performance was similarly effective or better than the Timed Up and Go and Four Square Step Test. This model could be used clinically to identify fall-risk individuals during a 6MWT, thereby finding people who were not intended for fall screening. Since model specificity was very high, the risk of accidentally misclassifying people who are a no fall-risk individual is quite low, and few people would incorrectly be entered into fall mitigation programs based on the test outcomes.