This study involves the functional testing of a new lower extremity prosthesis by healthy, active participants with fully healed transtibial (below knee) amputations. The study design calls for an experimental group of eleven participants who received two agonist-antagonist myoneural interfaces (AMIs) that were surgically constructed during a modified transtibial amputation procedure, and a control group of eleven matched participants who received standard transtibial amputations. The study protocol involves one or more of the following activities: 1. Collection of electromyography (EMG) data from participants' lower limbs to characterize muscle activation and create maps specific to individual participants, 2. Investigation of participants' capabilities to use a new lower extremity prosthesis that is designed to allow independent actuation of the ankle and subtalar joints, and offers EMG-modulated control over prosthetic joint position and stiffness, and 3. Exploration of AMIs as a means of communicating information between the participant and the new prosthesis using an experimental system involving EMG, functional electrical stimulation, and ultrasound. The hypothesis is that transtibial amputations involving AMIs can offer improved motor control of the new prosthesis while also enabling proprioceptive sensation (perception of the position, movement, and torque of the affected limb and prosthetic joint). The AMIs are expected to improve voluntary prosthetic control, improve prosthetic terrain adaptations, and offer new possibilities for bi-directional communication across the human-device interface.
BACKGROUND: Loss of limb profoundly impacts a person's health, productivity, independence, and quality of life. However, state-of-the-art medical and prosthesis technologies fall short of offering seamless human-device communication to those who require limb amputation. Ongoing, interactive efforts to advance amputation surgery techniques and develop novel "bionic" prostheses and prosthetic control systems are underway in an effort to address this interfacing challenge and thereby improve clinical outcomes within the population of amputees. We recently reported on the results of a first-in-human trial in which a prototype bionic prosthesis was tested in a recipient of a modified transtibial amputation. The modified amputation procedure involved the surgical construction of agonist-antagonist myoneural interfaces (AMIs) within the residual limb, where each AMI comprised two muscles - an agonist and an antagonist - connected in series. To enable the force produced by one muscle to cause stretch of its partner, "pulleys" were also constructed from the medial and lateral tarsal tunnels, including segments of each tunnel's native tendons, that were procured from the distal amputated limb and affixed to the residual limb tibia. The two AMIs were constructed via coaptation of the tibialis anterior and lateral gastrocnemius muscles to either end of the tendon portion passing through the proximally positioned tarsal tunnel, and coaptation of the tibialis posterior and peroneus longus muscles to either end of the tendon passing through the distally positioned tunnel. Following rehabilitation, this first recipient of the "AMI transtibial amputation" tested the feasibility of using his surgically constructed AMIs to control a prototype bionic prosthesis. The bionic prosthesis allowed motion in two degrees of freedom through independent actuation of powered ankle and subtalar joints, and the control algorithm allowed electromyography (EMG)-modulated control over prosthetic joint position and joint impedance. Functional testing involved linking the proximal and distal AMIs within the participant's residual limb to the prosthetic ankle and subtalar joints through the use of surface EMG electrodes, intramuscular fine-wire electrodes, and functional electrical stimulation. The results of performance testing in this first AMI recipient suggested that AMIs can provide a biological tissue interface that can potentially offer a person with an amputation intuitive motor control of the affected limb and a bionic prosthesis while also enabling proprioception. The unique biomimetic tissue architecture of the AMI recapitulates a dynamic, mechanically functional muscle-tendon-muscle linkage that inherently provides mechanoreceptive biological sensors. Consequently, the AMI tissue architecture inherently preserves natural, bi-directional communication between surgically reconstructed limb musculature and the central nervous system, thereby building on and offering advantages over previously described neural interfacing approaches such as targeted muscle reinnervation (TMR), regenerative peripheral nerve interfaces (RPNIs), and peripheral nerve interfaces. Additionally, the surgical design implemented in the AMI transtibial amputation preserves the native innervation and vascularization for each nerve and muscle component, thereby offering a more robust, viable surgical construction than either TMR or RPNI, which instead rely upon the less robust regenerative processes of reinnervation and revascularization for long term viability. Long term functionality of the AMI is facilitated by the incorporation of autologous tarsal tendon and tunnel components, eliminating the need for either allogeneic grafts or synthetic implant materials. By providing a platform for robust efferent decoding of movement intent, as well as usable afferent feedback from a prosthetic joint, the AMI transtibial amputation paradigm has the potential to reinstate the human central nervous system as the primary mediator of prosthetic joint control. STUDY OVERVIEW: The goal of this clinical trial is to evaluate the efficacy of the AMI transtibial amputation. STUDY POPULATION: This study calls for healthy, active participants with transtibial amputations: an experimental group of participants who received AMI transtibial amputations, and a control group of participants who received standard transtibial amputations. Each participant in the control group is prospectively matched to a participant in the experimental group to the degree possible based on time since amputation, body habitus, age, and biological sex. We anticipate that a planned number of thirty-two enrolled, consented participants will allow us to account for participant attrition over time. As the population of lower limb amputees consists of participants of all genders and ethnicities, and since it is not practical to attempt to match all aspects of this variation in the context of a small study, this study aims to reflect the variation in the population of amputees to the degree possible. EXPERIMENTAL SESSIONS: Biomedical data are collected from study participants in the Biomechatronics space within the MIT Media Lab in Cambridge, MA. Experimental group participants attend five or six sessions, with four sessions lasting approximately 4 hours each and the other one or two session(s) lasting up to 8 hours. Control group participants attend four sessions lasting approximately 4 hours each. HYPOTHESIS: Surgically constructed AMIs within the amputated residuum can afford an improved independent control of joint position and impedance in a multi-degree-of-freedom prosthesis while also reflecting proprioceptive sensation from each prosthetic joint onto the central nervous system. SPECIFIC AIM 1: Motion Control in Free Space Aim 1 investigates if AMIs can improve voluntary free-space prosthetic control. Experimental and control group participants' capabilities for prosthetic control are evaluated and compared based on EMG and biomechanical measurements obtained during free-space voluntary movement tasks. In Aim 1A, data are collected using surface EMG sensors to characterize muscle activation, create maps specific to individual participants, and inform sensor requirements for subsequent aims. The data are obtained from a large number of EMG sensors that are distributed over participants' lower limbs. The participant is asked to move the phantom and/or biological foot through the ankle and subtalar joint spaces during data collection. Aim 1B explores if AMIs can improve motion control of a prosthesis that allows independent actuation of powered ankle and subtalar joint motions and EMG-modulated control over prosthetic joint positions and stiffnesses. Surface EMG data are obtained using a small number of sensors that are placed on participants' lower limbs at locations informed by the results of Aim 1A. Joint state data are collected from sensors on the prosthesis and other noninvasive sensors. Participants are asked to move their phantom ankle joints, in some cases mirroring specified motions of their unaffected limbs, in order to control the prosthesis. Performance tasks include pointing the prosthetic foot toward a specified position and stiffening the prosthetic joint to hold that position for a specified time interval. SPECIFIC AIM 2: Terrain Adaptation Aim 2 determines if AMIs can improve voluntary and involuntary (reflexive) prosthetic terrain adaptations. Experimental and control group participants' capabilities for prosthetic control are evaluated and compared based on EMG, biomechanical, and kinematic measurements that are obtained as they walk and traverse various terrains in a motion capture space. Surface EMG data are obtained using a small number of sensors that are placed on participants' lower limbs at locations informed by the results of Aim 1. Biomechanical data are collected from sensors on the prosthesis and sensors embedded in the terrain equipment. Kinematic data are collected wirelessly using a twelve-camera motion capture system. To facilitate kinematic data collection, reflective markers are affixed to the participant's clothes or skin to enable visualization and tracking of anatomical landmarks. Participants perform level ground walking and terrain adaptation tasks. One task involves navigating an obstacle presented in the participant's path during level ground walking. The task involves eversion of the prosthetic subtalar joint such that the lateral edge of the prosthetic foot contacts a vertically offset block while the medial edge of the prosthetic foot remains at the base height. Other tasks involve either descending or ascending stairs in sequential steps. SPECIFIC AIM 3: Human-Device Communication Aim 3 explores if AMIs can enable new possibilities for bi-directional human-device communication and provides data toward developing closed-loop prosthetic control strategies. Experimental group participants' capabilities to control prosthetic motion and their associated proprioceptive perceptions are investigated. EMG, ultrasound, biomechanical, and psychometric data are collected in the presence of varying levels of functional electrical stimulation (FES). The FES delivers a periodic stream of electrical pulses to target muscles in the participant's affected and unaffected limbs, causing contraction. One performance task involves an experimental pedal-pushing set-up. The participant is blindfolded and asked to plantar flex the phantom ankle joint, which causes the prosthetic ankle joint to press down on a foot pedal against mechanical resistance. Participants are also asked to plantar flex at varying effort levels and, in accordance with prosthetic sensor data resulting from each effort level, FES is applied to specific target muscles. Another task involves applying FES to specific target muscles in the affected limb and asking the participant to describe the perceived motions and forces and mirror these in the unaffected limb. In addition to participant responses, data are collected from fine-wire electrodes, surface EMG and other noninvasive sensors including an ultrasound imaging probe, and sensors on-board the prosthesis. Fine-wire FES and EMG electrodes are included in this study to reduce crosstalk that would interfere with the implementation of the prosthetic control strategy. The fine wire electrodes are placed by an experienced clinician in an acute setting; they do not remain in the limb. FES settings are kept within historically safe limits at all times. The FES begins at low intensity and is slowly increased until either the participant reports that a limit of comfortable stimulation has been reached or the historically safe limit is reached. The lower of these two values is established as a hard-stop reference for the FES setting.
Study Type
INTERVENTIONAL
Allocation
NON_RANDOMIZED
Purpose
BASIC_SCIENCE
Masking
NONE
Enrollment
14
Two Agonist-antagonist myoneural interfaces (AMIs) were surgically constructed during a modified transtibial amputation procedure. Each AMI was made of natively innervated and vascularized muscle segments - an agonist and antagonist - that were surgically connected in series within the amputated residuum. Tarsal tunnels, including segments of each tunnel's native tendon component, were procured from the amputated joint. The tunnels were affixed to the residual limb tibia and the AMIs were constructed by coaptation of an agonist and an antagonist muscle to either end of the tendon passing through the tunnel. Consequently, the force produced by one muscle stretches its partner such that the AMI can communicate signals from the mechanoreceptors in both muscles to the central nervous system.
A standard transtibial amputation was performed according to traditional techniques. No surgical construction of agonist-antagonist myoneural interfaces (AMIs) was performed.
Massachusetts Institute of Technology
Cambridge, Massachusetts, United States
Stability of Joint Position Control in Free Space
The stability of joint position control in free space is quantified by the number of the distinct synergy activations (distinct movements) achieved out of a total of four targeted movements of interest: (1) ankle plantar flexion (toe down), (2) dorsiflexion (toe up), (3) subtalar joint eversion (sole of foot outward), and (4) subtalar joint inversion (sole of foot inward). For each movement, the subject is asked attempt the movement while the distinct synergy activation/neural signals are quantified using electromyography (EMG) data. An outcome of 4 indicates that the subject was able to produce distinct activations for each of the 4 targeted movements. An outcome of less than 4 indicates that although a subject attempted the movement, they were not able to successfully produce distinct activations for some portion of the targeted movements.
Time frame: 1 time point, post-amputation
Economy of Motion for Free Space Movements
The economy of motion is computed as the total travel distance through synergy space, normalized by the minimum possible/most direct travel path, to reflect control efficiency. Given this definition, the economy of motion indicates the trajectory straightness of movements that were produced to achieve the target discrete movements. For this study, the movements were ankle plantar-dorsiflexion and subtalar inversion-eversion. An outcome of 100% represents how the two movements together could allow for an economy of the targeted movements in that space, indicating perfect economy of motion. The percentage may decrease if a subject achieves the targeted movements in a less efficient manner. For these movements, the economy of motion was evaluated under increasing time constraints from 2.0 s to 1.5 s, 1 s, 0.8 s, and 0.5 s.
Time frame: 1 time point, post-amputation
Late Swing Ankle Plantar Flexion During Stair Descent
To address the clinical trial aim of determining whether AMIs can improve prosthetic terrain adaptations, we assessed swing phase control during stair descent by measuring the capability of the neuroprosthesis to exhibit prosthetic ankle joint plantar flexion characteristic of stair descent. This metric was defined as the change in ankle joint angle from terminal stance to terminal swing, capturing the user's ability to distinctly control joint angle transitions across gait phases of stair descent. For further details see: H. Song, T.-H. Hsieh, S. H. Yeon, T. Shu, M. Nawrot, C. F. Landis, G. N. Friedman, E. A. Israel, S. Gutierrez-Arango, M. J. Carty, L. E. Freed, H. M. Herr, Continuous neural control of a bionic limb restores biomimetic gait after amputation. Nat Med 30, 2010-2019 (2024).
Time frame: 1 time point, post-amputation
Late Swing Ankle Dorsiflexion During Stair Ascent
To address the clinical trial aim of determining whether AMIs can improve prosthetic terrain adaptations, we assessed swing phase control during stair ascent by measuring the capability of the neuroprosthesis to exhibit prosthetic ankle joint dorsiflexion characteristic of stair ascent. This metric was defined as the change in ankle joint angle from terminal stance to terminal swing, capturing the user's ability to distinctly control joint angle transitions across gait phases of stair ascent. For further details see: H. Song, T.-H. Hsieh, S. H. Yeon, T. Shu, M. Nawrot, C. F. Landis, G. N. Friedman, E. A. Israel, S. Gutierrez-Arango, M. J. Carty, L. E. Freed, H. M. Herr, Continuous neural control of a bionic limb restores biomimetic gait after amputation. Nat Med 30, 2010-2019 (2024).
Time frame: 1 time point, post-amputation
Correlation of Ankle Joint Proprioception
Correlation of Ankle Joint Proprioception was measured by applying functional electrical stimulation (FES) to either the tibialis anterior (TA) or lateral gastrocnemius (LG) and instructing the subject to resist the experienced movement. The level of correlation (R2) for both dorsiflexion and plantarflexion between the applied stimulation current and the subjects perceived joint counter-torque for the intervention subjects was compared to control subjects.
Time frame: 1 time point, post-amputation
Controllability Over Prosthetic Joint Dorsi and Plantar Flexion
We analyzed the correlation (R2) between agonist contraction and antagonist stretch during dorsi and plantar flexion movements. Muscle fascicle strains were estimated from ultrasound data recorded from the agonist and antagonist muscles (e.g. lateral gastrocnemius (LG) and tibialis anterior (TA) during plantar flexion). Fascicle measurements were generated via optical tracking software. In the AMI subjects, because of the mechanical coupling of the agonist and antagonist muscles within the residual limb, a volitional contraction of the agonist muscle inherently induces a passive stretch in the corresponding antagonist muscle. On the other hand, in the control subjects, because the residual muscles are not coapted, we expected to find limited stretch in the antagonist muscle during agonist contraction. With limited stretch in the antagonist from an agonist contraction, perceived joint movement would be lowered from antagonist stretch afferents.
Time frame: 1 time point, post-amputation
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.