Automatic vs hand-controlled walking of paraplegics

A rule-based control and its application in functional electrical stimulation (FES) assisted walking of subjects with paraplegia are described in this paper. The design of rules for control comprises the following two steps: (1) determination of muscle activation patterns by using a fully customized spatial (3D) model of paraplegic walking, and (2) learning of rules, that is, correlation between the muscle activation patterns and kinematics of walking by means of an artificial neural network. The adopted FES system activated eight muscle groups with surface electrodes. The only joints allowing movement in the coronal plane were the hips, and externally controlled joints in sagittal plane were ankles, knees and hips. The simulation minimized the tracking error of the joint angles and the total activation of all eight muscles being stimulated. A radial-basis function artificial neural network was applied for learning of rules. Three automatically controlled modes (slow, near-normal, and near-ballistic) and hand-controlled walking were evaluated in six subjects with a complete spinal cord lesion (T8-T10). The performance of walking was assessed by the following: (1) energy consumption based on oxygen uptake, (2) physiological cost index, (3) maximum speed of walking, and (4) a questionnaire. The results showed that all modes of walking are achievable and that automatic control leads to more efficient and faster walking. The speed of walking achieved by automatic control was almost three times bigger compared with the speed of hand-controlled walking. The energy cost and rate decreased significantly when automatic control was applied; yet, they were still much bigger than the values measured in able-bodied subjects. The objective outcome measures suggest that the near-ballistic walking was the most effective, yet a questionnaire shows that most subjects preferred slow walking. The most likely reason for the preference of lower efficiency walking over the faster end energy efficient near-ballistic walking was that paraplegic patients had difficulties in synchronizing the voluntary movement of the trunk and arms to the artificially controlled movements of legs.     

服务于功能性电刺激的双足步态实验

背景：功能性电刺激利用低频弱电流脉冲刺激失去神经控制的肌肉已经在截瘫行走的临床应用中取得了小范围成功,但现有的电刺激模式存在不灵活、不易操作、且稳定性不高的缺点。 目的：基于步态分析方法,研究涉及到步行动作的各肌肉群的协同动作关系,将肌肉电刺激模式简化为无需患者操作的规律性控制策略,并验证该策略在功能性电刺激实验中的有效性。 方法：针对双足步行的特点,提出一种基于关节角变化趋势及肌电信号强度变化的步态研究方法,旨在服务于功能性电刺激的设计,为下肢肌肉提供理想的电刺激模式,使人体产生相应的肌肉群协同动作,从而使受试者最终实现非自主控制的行走运动。 结果与结论：实验结果一方面验证了基于步态分析的电刺激模式设计是可行的,对今后加入更复杂的控制方式提供了依据,另一方面也为未来开展的瘫痪患者临床康复实验研究打下了基础。     

Analysis of Musculoskeletal Loadings in Lower Limbs During Stilts Walking in Occupational Activity

Construction workers often use stilts to raise them to a higher level above ground to perform many tasks, such as taping and sanding on the ceiling or upper half of a wall. Some epidemiological studies indicated that the use of stilts may place workers at increased risk for knee injuries or may increase the likelihood of trips and falls. In the present study, we developed an inverse dynamic model of stilts walking to investigate the effects of this activity on the joint moments and musculoskeletal loadings in the lower limbs. The stilts-walk model was developed using the commercial musculoskeletal simulation software AnyBody (version 3.0, Anybody Technology, Aalborg, Denmark). Simulations were performed using data collected from tests of four subjects. All subjects walked without or with stilts through a 12-m straight path. The moments of the knee, hip, and ankle joints, as well as forces in major muscles or muscle groups in the lower limbs, for stilts walking were compared with those for normal walking. Our simulations showed that the use of stilts may potentially increase the peak joint moment in knee extension by approximately 20%; induce 15% reduction and slight reduction in the peak joint moments in ankle plantar flexion and hip extension, respectively. The model predictions on the muscle forces indicated that the use of stilts may potentially increase loadings in five of eight major muscle groups in the lower extremities. The most remarkable was the force in rectus femoris muscle, which was found to potentially increase by up to 1.79 times for the stilts walking compared to that for the normal walking. The proposed model would be useful for the engineers in their efforts to improve the stilts design to reduce musculoskeletal loadings and fall risk.     

基于中枢模式发生器控制的电刺激步行康复系统设计与实验

研发功能性电刺激(FES)康复系统来实现下肢瘫痪病人的行走运动。控制器采用基于中枢模式发生器(CPG)的仿生控制机理。针对基于递归神经振荡器的CPG模型,研究关键参数(激励性输入、时间常数、感觉反馈、输出阈值)对CPG输出幅值、频率、相位以及占空比的影响。建立包含12个神经元的CPG网络,以控制双腿的4个关节(左/右髋关节和膝关节)和八组肌群(左/右髂腰肌、臀大肌、股直肌和腘绳肌)。所搭建的实验系统和平台,包括悬吊减重系统、助行系统、电刺激系统和运动检测系统。在正常受试者参与的实验中,受试者在CPG控制的电刺激作用下产生非自主行走运动。实验结果表明,受试者双腿的髋关节和膝关节角度与正常人自主行走时的数据达到定性吻合,验证所设计的CPG控制器在FES康复系统中的可行性与有效性。     

功能性电刺激技术在截瘫行走中的应用研究进展

截瘫是由脊髓损伤所造成的双下肢严重残疾,近年来的发病率呈显著上升趋势。截瘫最主要的病症就是行走能力的丧失。经过40多年的研究表明,功能性电刺激技术能成功地恢复截瘫患者的部分运动功能,是现代康复工程领域很有应用前景的一项新技术,正在受到越来越多的重视。本文专门针对用于截瘫行走的功能性电刺激技术,介绍了与之相关的背景知识和研究进展。     

踝足矫形器人机耦合仿真刚度对步态生物力学影响的评价方法

目的 为更好评价踝足矫形器(ankle-foot orthosis, AFO)的人机工效,提出一种 AFO 人机耦合仿真刚度对步态生物力学影响的评价方法。 方法 首先,评测两种 AFO 的力学特性并量化其刚度;其次,采集 30 名受试者在正常与穿戴两种不同刚度 AFO 行走条件下的运动学及动力学数据;最后,通过仿真计算,定量对比分析 AFO 刚度对 行走中下肢关节角度、关节力矩与肌肉力的影响。 结果 在穿戴两种刚度 AFO 情况下,支撑相踝关节背屈峰值角度、膝关节屈曲峰值角度和髋关节伸展峰值角度均显著增加(P<0. 05),比目鱼肌与腓肠肌峰值肌肉力呈现增加趋势。 踝关节跖屈峰值角度、髋关节屈曲峰值角度与髋关节屈曲峰值力矩均显著降低(P<0. 05)。 结论 本文所提出的 AFO 人机耦合仿真方法可以有效实现不同刚度对步态生物力学影响的定量评估,该方法同样也适用于评估和优化其他辅助器具的人机工效,可以为 AFO 的选配、适配、优化设计提供方法指导。     

The effect of ageing in spinal cord injured humans on the blood pressure and heart rate responses during fatiguing isometric exercise

Groups of 50 healthy male controls and 50 subjects suffering from paraplegia (aged 20–65 years) were examined as to the inter-relationships between age, paraplegia and the strength, endurance, blood pressure and heart rate responses to fatiguing isometric exercise. Contractions were maintained in both groups under voluntary effort and through a contraction induced by electrical stimulation in the paraplegic group. All contractions were maintained to fatigue at a tension of 40% of the maximal muscle strength in either the handgrip or quadriceps muscles. Muscle strength of the handgrip was higher in the paraplegic subjects than in the controls, averaging 589 N and 463 N, respectively for the two groups. In contrast, quadriceps leg extension strength averaged 696 N in the controls and 190 N in the paraplegic groups; for both groups, ageing was associated with a reduction in muscle strength. While leg endurance was less in the paraplegic group than the control group, handgrip endurance was similar in the two groups, endurance increasing with ageing in both the controls and paraplegics. Both systolic and diastolic blood pressures increased at rest in paraplegic and control subjects with age. The magnitude of the pressor response to exercise also increased with age. This was true during both voluntary exercise and exercise induced through electrical stimulation in the paraplegic groups. The heart rate response (change in heart rate during exercise) to a fatiguing isometric handgrip contraction decreased by about 50% between the ages of 20 and 60 years in both the controls and paraplegics for isometric handgrip exercise. In contrast, heart rate changed little with age during contractions of the quadriceps muscle in paraplegics which were induced by electrical stimulation. Electronic Publication     

Walking cycle recording and analysis for FNS-assisted paraplegic walking

In closed-loop functional neuromuscular stimulation (FNS)-assisted paraplegic walking, there is a need for reference leg motion trajectories that describe the desired walking cycle. These reference trajectories were defined as the angular changes between the leg segments, as measured by an electrogoniometer system. For each leg, the hip, knee and ankle trajectories of normal individuals during slow walking were measured and sampled over a number of cycles. Additionally, foot contact with the ground was measured to synchronise the various walking trajectories. Each joint's angular trajectory was averaged over a number of walking cycles, using an interpolation method based on a discrete Fourier transform (DFT) and inverse DFT technique, to expand all the signals to the same length. In this way an average walking cycle was obtained for each trial, representing the six averaged leg motion trajectories for one walking cycle. Angular trajectories and walking parameters for slow and normal walking were compared so as to investigate principles of walking cycle adaptation necessary to stabilise the body during slow walking. In general, angular trajectories were similar for different subjects, but different for different walking speeds, due to the greater demands on maintaining stability during slow walking. It can be concluded that normal speed walking consists of separate, unstable phases, whereas slow speed walking, relevant for paraplegic walking, requires stabilising each separate phae of the walking cycle.     

An artificial reflex improves the perturbation-resistance of a human walking simulator

Most walking assist systems reported are not available for real-world environment, where frequent perturbations are caused by slips, uneven terrain, slopes and obstacles. On the other hand, it is evident that human beings cope with those perturbations, especially when the perturbations cannot be predicted or perceived in advance, with reflexes, which cause relatively fixed muscular responsive patterns to perturbations unconsciously within a short period of time ranging from several 10 to 200 ms. Our ultimate goal is to realize artificial reflexes in real-world walking support systems for those paralyzed people, whose afferent and efferent neural pathways are usually weakened, so that the reflexive system is also impaired to a certain degree. This goal needs both qualitative and quantitative understanding of human reflexive mechanism during walking. However, except for some hypotheses about the underlying neural mechanisms of the reflexes during walking, there is no widely accepted unified theory, nor are there clear experimental results that could be directly quoted in the disciplines of physiology and motor control. Our approach includes (1) acquiring muscle activity profiles during normal walking and slip-perturbed walking by recording and processing Electromyographic (EMG) signals of several walking-related muscles, in human gait experiments; (2) developing a central-pattern-generator (CPG) based neuro-musculo-skeletal simulation model; (3) comparing joint trajectories of the simulation model with those of a human subject during normal walking to verify the simulation model’s conformity with human walking; (4) using muscle activity profiles of reflexive responses to slip-perturbation during walking to construct a rapid responding pathway. The results showed that, (1) The simulation model could show behavior resembling that of normal human walking; (2) in the case of occurrence of slip-perturbation, the rapid responding pathway could improve the perturbation-resistance and maintain the balance for the walking; (3) using the simulation model, several hypotheses on underlying neuro-mechanism were investigated. These reveal the possibility to realize the artificial reflex for the paralyzed people.     

A decentralized adaptive fuzzy robust strategy for control of upright standing posture in paraplegia using functional electrical stimulation

In this paper, we present a novel decentralized robust methodology for control of quiet upright posture during arm-free paraplegic standing using functional electrical stimulation (FES). Each muscle–joint complex is considered as a subsystem and individual controllers are designed for each one. Each controller operates solely on its associated subsystem, with no exchange of information between them, and the interaction between the subsystems are taken as external disturbances. In order to achieve robustness with respect to external disturbances, unmodeled dynamics, model uncertainty and time-varying properties of muscle–joint dynamics, a robust control framework is proposed. The method is based on the synergistic combination of an adaptive nonlinear compensator with sliding mode control (SMC). Fuzzy logic system is used to represent unknown system dynamics for implementing SMC and an adaptive updating law is designed for online estimating the system parameters such that the global stability and asymptotic convergence to zero of tracking errors is guaranteed. The proposed controller requires no prior knowledge about the dynamics of system to be controlled and no offline learning phase. The results of experiments on three paraplegic subjects show that the proposed control strategy is able to maintain the vertical standing posture using only FES control of ankle dorsiflexion and plantarflexion without using upper limbs for support and to compensate the effect of external disturbances and muscle fatigue.     

Cycle-to-cycle control of swing phase of paraplegic gait induced by surface electrical stimulation

Parameterised swing phase of gait in paraplegics was obtained using surface electrical stimulation of the hip flexors, hamstrings and quadriceps; the hip flexors were stimulated to obtain a desired hip angle range, the hamstrings to provide foot clearance in the forward swing, and the quadriceps to acquire knee extension at the end of the swing phase. We report on two main aspects; optimisation of the initial stimulation parameters, and parameter adaption (control). The initial stimulation patterns were experimentally optimised in two paraplegic subjects using a controlled stand device, resulting in an initial satisfactory swinging motion in both subjects. Intersubject differences appeared in the mechanical output (torque joint) per muscle group. During a prolonged open-loop controlled trial with the optimised but unregulated stimulation onsets and burst duration for the three muscle groups, the hip angle range per cycle initially increased above the desired value and subsequently decreased below it. The mechanical performance of the hamstrings and quadriceps remained relatively unaffected. A cycle-to-cycle controller was then designed, operating on the basis of the hip angle ranges obtained in previous swings. This controller successfully adapted the burst duration of the hip flexors to maintain the desired hip angle range.     

Feedback control of unsupported standing.

This paper presents the results of continuing work on feedback control of unsupported standing in paraplegia. Our experimental setup considers a situation in which all joints above the ankle are braced, and stabilising torque at the ankle is generated by stimulation of the plantarflexors. A previous study showed that short periods of unsupported standing with paraplegic subjects could be achieved. In order to improve consistency and reliability of unsupported standing we are currently investigating several modifications to the control strategy. The paper reports progress towards this goal.     

截瘫FES行走过程中空间平衡性的监护新方法

首次提出了以空间危势轨迹图(SRTT)来评估和监护截瘫功能性电刺激(FES)行走空间平衡性的新方法。其主要构件是一套基于标准步行器的测试系统,该系统可以实时获得上肢支撑力数据,通过上躯干力学模型的重心定位结合步行器倾翻指数定义转化为SRTT,描述行走过程中各轴向空间的平衡状况。相关实验和临床测试结果显示新方法具有准确快速的特点,在行走能力评估和监护方面有很好的临床应用前景,有望为下一步制定针对FES有效使用的相应国家级康复标准奠定基础。     

Comparison of stimulation patterns for FES-cycling using measures of oxygen cost and stimulation cost

AIM: The energy efficiency of FES-cycling in spinal cord injured subjects is very much lower than that of normal cycling, and efficiency is dependent upon the parameters of muscle stimulation. We investigated measures which can be used to evaluate the effect on cycling performance of changes in stimulation parameters, and which might therefore be used to optimise them. We aimed to determine whether oxygen cost and stimulation cost measurements are sensitive enough to allow discrimination between the efficacy of different activation ranges for stimulation of each muscle group during constant-power cycling. METHODS: We employed a custom FES-cycling ergometer system, with accurate control of cadence and stimulated exercise workrate. Two sets of muscle activation angles ("stimulation patterns"), denoted "P1" and "P2", were applied repeatedly (eight times each) during constant-power cycling, in a repeated measures design with a single paraplegic subject. Pulmonary oxygen uptake was measured in real time and used to determine the oxygen cost of the exercise. A new measure of stimulation cost of the exercise is proposed, which represents the total rate of stimulation charge applied to the stimulated muscle groups during cycling. A number of energy-efficiency measures were also estimated. RESULTS: Average oxygen cost and stimulation cost of P1 were found to be significantly lower than those for P2 (paired t-test, p<0.05): oxygen costs were 0.56+/-0.03l min-1 and 0.61+/-0.04l min-1 (mean+/-S.D.), respectively; stimulation costs were 74.91+/-12.15 mC min-1 and 100.30+/-14.78 mC min-1 (mean+/-S.D.), respectively. Correspondingly, all efficiency estimates for P1 were greater than those for P2. CONCLUSION: Oxygen cost and stimulation cost measures both allow discrimination between the efficacy of different muscle activation patterns during constant-power FES-cycling. However, stimulation cost is more easily determined in real time, and responds more rapidly and with greatly improved signal-to-noise properties than the ventilatory oxygen uptake measurements required for estimation of oxygen cost. These measures may find utility in the adjustment of stimulation patterns for achievement of optimal cycling performance.     

Classifying Lower Extremity Muscle Fatigue During Walking Using Machine Learning and Inertial Sensors

Fatigue in lower extremity musculature is associated with decline in postural stability, motor performance and alters normal walking patterns in human subjects. Automated recognition of lower extremity muscle fatigue condition may be advantageous in early detection of fall and injury risks. Supervised machine learning methods such as support vector machines (SVMs) have been previously used for classifying healthy and pathological gait patterns and also for separating old and young gait patterns. In this study we explore the classification potential of SVM in recognition of gait patterns utilizing an inertial measurement unit associated with lower extremity muscular fatigue. Both kinematic and kinetic gait patterns of 17 participants (29 ± 11 years) were recorded and analyzed in normal and fatigued state of walking. Lower extremities were fatigued by performance of a squatting exercise until the participants reached 60% of their baseline maximal voluntary exertion level. Feature selection methods were used to classify fatigue and no-fatigue conditions based on temporal and frequency information of the signals. Additionally, influences of three different kernel schemes (i.e., linear, polynomial, and radial basis function) were investigated for SVM classification. The results indicated that lower extremity muscle fatigue condition influenced gait and loading responses. In terms of the SVM classification results, an accuracy of 96% was reached in distinguishing the two gait patterns (fatigue and no-fatigue) within the same subject using the kinematic, time and frequency domain features. It is also found that linear kernel and RBF kernel were equally good to identify intra-individual fatigue characteristics. These results suggest that intra-subject fatigue classification using gait patterns from an inertial sensor holds considerable potential in identifying “at-risk” gait due to muscle fatigue.     

Assessment of spatio-temporal parameters during unconstrained walking

This paper presents an analysis of spatio-temporal gait parameters during overground walking based upon a method that needs only lower trunk accelerations. Twenty-six healthy young subjects and 15 healthy elderly subjects participated in an experiment where overground walking was studied at different speeds. Accelerations of the lower trunk were measured by a tri-axial accelerometer connected to a portable data logger carried on the body. An analysis of trunk acceleration data produced temporal gait parameters (duration of subsequent stride cycles and left/right steps) and convincing estimations of spatial parameters (step length and walking speed). Typical differences in spatio-temporal gait parameters between young and elderly subjects could be demonstrated, i.e. a limited range of walking speeds, smaller step lengths, and a somewhat higher variability of temporal parameters in elderly subjects. It is concluded from these results that essential spatio-temporal gait parameters can be determined during overground walking using only one tri-axial accelerometer. The method is easy-to-use and does not interfere with regular walking patterns. Both the accelerometer and the data logger can be miniaturised to one small instrument that can be carried on the trunk during hours of walking. Thus, the method can easily be incorporated in current activity monitors so that 24-h monitoring of postures and activities can be combined with assessment of gait characteristics during these monitoring periods. In addition, the presented method can be a basis for more sophisticated gait analyses during overground walking, e.g. an analysis of kinematic signals or muscle activity within subsequent stride cycles.     

Relationships between muscle mitochondrial DNA content, mitochondrial enzyme activity and oxidative capacity in man: alterations with disease

Muscle mitochondrial content is tightly regulated, and requires the expression of both nuclear and mitochondrial genes. In addition, muscle mitochondrial content is a major determinant of aerobic exercise capacity in healthy subjects. The current study was designed to test the hypothesis that in healthy humans, muscle mitochondrial DNA (mtDNA) content is correlated with citrate synthase activity (a representative nuclear-encoded mitochondrial enzyme) and aerobic exercise capacity as defined by whole-body peak oxygen consumption (V˙O₂). Furthermore, it was postulated that these relationships might be altered with disease. Twelve healthy and five paraplegic subjects underwent exercise testing and vastus lateralis muscle biopsy sampling. An additional ten healthy subjects and eight patients with unilateral peripheral arterial disease (PAD) underwent exercise testing and gastrocnemius muscle biopsy sampling. Citrate synthase activity and mtDNA content were positively correlated in the vastus lateralis muscles from the healthy subjects. This relationship was similar in muscle from paraplegic subjects. mtDNA content was positively correlated with peak V˙O₂ in the healthy subjects and in the paraplegic subjects in whom peak V˙O₂ had been elicited by functional electrical stimulation of the muscle. In contrast, the PAD subjects demonstrated higher mtDNA contents than would have been predicted based on their claudication-limited peak V˙O₂. Thus, in healthy humans there are strong relationships between muscle mtDNA content and both muscle citrate synthase activity and peak V˙O₂. These relationships are consistent with coordinant nuclear DNA and mtDNA expression, and with mitochondrial content being a determinant of aerobic exercise capacity. The relationships seen in healthy humans are quantitatively similar in paraplegic patients, but not in patients with PAD, a disease which is associated with a metabolic myopathy. The relationships between mtDNA content, mitochondrial enzyme activities and exercise capacity provide insight into the physiologic and pathophysiologic regulation of muscle mitochondrial expression.     

踝关节外骨骼助行模式对下肢肌肉激活与协调模式的影响EI北大核心CSCD

踝关节外骨骼可用来提高人的行走效率,辅助老年人、运动功能障碍患者等进行日常活动或康复训练,但外骨骼的助行模式会对穿戴者的下肢肌肉激活与协调模式产生影响。本文利用一款绳驱动踝关节外骨骼,设计了不同助力时机与助力大小组合的助行模式,采集了7名穿戴者在跑步机上以1.25 m/s速度水平行走时的下肢表面肌电信号,研究不同助行模式对穿戴者下肢肌肉激活与协调模式的影响。实验结果表明,比目鱼肌激活程度在踝关节外骨骼助力时有明显降低,在助力时机为步态周期49%、助力大小为0.7 N·m/kg时,最高可降低(38.5±10.8)%。并且,相对于助力时机,助力大小对比目鱼肌激活程度影响更为显著。踝关节外骨骼不同模式助行时,所测量下肢肌肉可分解为5个基本协同模式,且合适的助力时机与助力大小条件下,下肢肌肉协调模式和正常行走相比改变较小。此外,比目鱼肌-胫骨前肌、股直肌-半腱肌的协同收缩度在外骨骼助力时比正常行走均有升高。本研究有助于理解健康穿戴者如何调整自身的神经肌肉控制机制来适应不同外骨骼助力,并为选择合适的助行模式以及合理利用外骨骼提高行走效率提供依据。     

Modulation of human cutaneous reflexes during rhythmic cyclical arm movement

The organization and pattern of cutaneous reflex modulation is unknown during rhythmic cyclical movements of the human upper limbs. On the assumption that these cyclic arm movements are central pattern generator (CPG) driven as has been suggested for leg movements such as walking, we hypothesized that cutaneous reflex amplitude would be independent of electromyographic (EMG) muscle activation level during rhythmic arm movement (phase-dependent modulation, as is often the case in the lower limb during locomotion). EMG was recorded from eight muscles crossing the human shoulder, elbow, and wrist joints while whole arm rhythmic cyclical movements were performed. Cutaneous reflexes were evoked with trains of electrical stimulation delivered at non-noxious intensities (approximately 2 x threshold for radiating paresthesia) to the superficial radial nerve innervating the lateral portion of the back of the hand. Phasic bursts of rhythmic muscle activity occurred throughout the movement cycle. Rhythmic EMG and kinematic patterns were similar to what has been seen in the human lower limb during locomotor activities such as cycling or walking: there were extensive periods of reciprocal activation of antagonist muscles. For most muscles, cutaneous reflexes were modulated with the movement cycle and were strongly correlated with the movement-related background EMG amplitude. It is concluded that cutaneous reflexes are primarily modulated by the background muscle activity during rhythmic human upper limb movements, with only some muscles showing phase-dependent modulation.