Ultrasound Imaging Based Sensing of Human Ankle Motion Intent and Control Strategies for Ankle Assistance

Ultrasound Imaging Based Sensing of Human Ankle Motion Intent and Control Strategies for Ankle Assistance

Ultrasound Imaging Based Sensing to Predict Human Ankle Movement Intent and Assist-As-Needed Control Using Functional Electrical Stimulation and Powered Ankle Exoskeleton

Robotic therapies aim to improve limb function in individuals with neurological injury. Modulation of robotic assistance in many of these therapies is achieved by measuring the extant volitional strength of limb muscles. However, current sensing techniques, such as electromyography, are often unable to correctly measure the voluntary strength of a targeted muscle. The difficulty is due to their inability to remove ambiguity caused by interference from activities of neighboring muscles. These discrepancies in the measurement can cause the robot to provide inadequate assistance or over-assistance. Improper robotic assistance slows function recovery, and can potentially lead to falls during robot-assisted walking. An ultrasound imaging approach is an alternative voluntary strength detection methodology, which can allow direct visualization and measurement of muscle contraction activities. The aim is to formulate an electromyography-ultrasound imaging-based technique to sense residual voluntary strength in ankle muscles for individuals with neuromuscular disorders. The estimated voluntary strength will be involved in the advanced controller's design of robotic rehabilitative devices, including powered ankle exoskeleton and functional electrical stimulation system. It is hypothesized that the ankle joint voluntary strength will be estimated more accurately by using the proposed electromyography-ultrasound imaging-based technique. And this will help the robotic rehabilitative devices achieve a more adaptive and efficient assistance control, and maximize the ankle joint rehabilitation training benefits.

The study will recruit up to 20 persons without any neuromuscular diseases and 5 persons with incomplete spinal cord injury (iSCI) or transverse myelitis, who are 18-64 years old for experiments. Research activities include an average of 8 experimental sessions. The sessions can be more or less than 8 sessions depending on subject availability, sufficient data for analysis, or need to obtain more data for the analysis. The first aim is to investigate ultrasound measurements of muscle movements and muscle electrical activity (sEMG) to predict the voluntary motion or torque produced at the human ankle joint. Based on the voluntary motion or torque, functional electrical stimulation (FES) will be used to artificially stimulate the ankle muscles or a powered ankle exoskeleton will be used to provide additional assistance at the ankle joint. FES electrodes will be placed at appropriate locations at the skin surface of the lower limb such that the most optimal motion is obtained. Task Type 1: Voluntary sitting state tasks without and with external assistance. Session 1: Voluntary ankle up and ankle down motion without FES. A pair of sEMG sensors will be placed to the participant's lower leg front and back muscles through double-sided tapes. A pair of US imaging transducers will be attached to the same lower-leg front and back muscles to image the muscle's activities during the ankle joint's up and down motion. A pair of movement sensors will be attached to the shank and foot to measure the angular position and velocity of the ankle joint. A force sensor will be placed between the foot sole and a rigid flat platform to measure the interaction force during the ankle up and ankle down motion. US imaging measurements, the electrical signals from sEMG sensors, the contact force from the force sensor, and the movement sensor and camera measurements will be collected simultaneously. Data may be collected from one or both of a participant's legs. Session 2: Voluntary ankle side motion towards inside and outside without FES. During session 2, the participant will perform side-to-side ankle movements. A similar training procedure, as in session 1, will be prescribed before the actual experiments. Multiple measurements from the sensors (same as in session 1) attached to the participant's leg and foot will be obtained during the side-to-side ankle motion. Both in session 1 and session 2, for each motion direction (up, down, inside, and outside), 10 repeated trials will be run, and each trial will last for 10 mins including training, data collection, data saving, and rest period to avoid muscle fatigue. Each session may last approximately up to 4 hours. Session 3: Assistive ankle up and down motion with the FES. A similar configuration as mentioned in session 1 will be applied in this session. Additionally, the FES will be used to control the ankle up or down a motion by taking the US imaging-based measurements as feedback. In the FES system, for ankle up motion, the negative electrode will be placed over the muscle belly of the front muscle, while the positive electrode will be placed further down the shank. For ankle down motion, two large electrodes will be used to allow activation of the whole back muscles. These electrodes will be connected to an electrical stimulator, which is controlled via computer software. For both ankle up and ankle down motion, a low current waveform will be used with 100-400 microseconds pulse width and 20-40 Hz frequency. The current range will be between 0-60 mA. In all cases, the pulse duration defined for use will not exceed the upper limit of 400 microseconds. The stimulation parameters used for a participant will be documented for each experimental session and any changes in the stimulation parameters during a session will also be recorded. Session 4: Assistive ankle side motion towards inside and outside with the FES system. A similar configuration as mentioned in session 2 will be applied in this session. The FES will be used to artificially stimulate the muscles and control ankle side motion towards inside or outside by taking the US imaging-based measurements as feedback. For both ankle side motion towards inside and outside situations, the same FES parameters and modulation method will be implemented as described in session 3. During the experiment procedures, the US imaging-based measurements and the movement and camera measurements will be collected simultaneously. Data may be collected from one or both of the participants' legs. Both in session 3 and session 4, for each motion direction (up, down, inside, and outside), 10 repeated trials will be run, and each trial will last for 10 mins including training, data collection, data saving, and rest period to avoid muscle fatigue. Each session may last approximately up to 4 hours. The participant may be asked to come back if the data obtained during a session is not found to be satisfactory during analysis. The next visit will be conducted at the participant's convenience and at least after 24 hours for session 1 and 2, as well as 72 hours for session 3 and 4. Task Type 2: Voluntary dynamic walking tasks on a treadmill without and with external assistance Session 5: Dynamic walking task without FES or ankle exoskeleton. Similar to session 1 and session 3 in Task Type 1, the sEMG sensors and ultrasound imaging transducers will be attached to the same positions on the front and back shank mentioned before. Multiple reflective markers will be attached to the subject's lower limbs to measure the kinematics data based on a motion capture system. During session 5, the participant will be asked to stand still for 30 seconds on the treadmill with two split belts. The force sensors below each belt will be used to measure the ground reaction force between the belt and the participant's feet soles, and at the same time, the sEMG signals and ultrasound imaging of the lower leg muscles will also be collected during the static standing posture. Then the participant will be asked to walk on the treadmill with 0.5, 0.75, 1.0, 1.25, and 1.5 m/s. For each speed, the first 5 minutes of walking will be given as a warmup for the participant to get used to the walking pattern with the wearable. After the warmup, in the following 2 minutes for each walking speed, the ground reaction force, sEMG signals and ultrasound imaging of the lower leg muscles, and the coordinates of the reflective markers will be recorded synchronously and simultaneously. Session 6: Dynamic walking task with FES or ankle exoskeleton but no powered assistance. Apart from the setup described in session 1, the FES electrodes will be placed on the front shank muscle and the ankle exoskeleton will be placed on the back shank. The same data collection as mentioned in session 5 will also be included. In this session, there will be no stimulation on the participant's skeletal muscles and no power for the ankle exoskeleton motor. Session 7: Dynamic walking task with FES or ankle exoskeleton based on an old control approach. The experimental setup and procedures will be kept the same as mentioned in session 2. During the walking dynamic task, the powered FES and ankle exoskeleton will be used to provide assistance for toe up and toe down motions according to human movement intent based on a prevalently applied control approach in other publications. Throughout each of the 2 minutes data collection duration, the sEMG signals will be collected and processed in real-time to control the FES or ankle exoskeleton system. In addition, the time sequence of ultrasound imaging from the front and back shank muscles, the reflective markers coordinates, the ground reaction force, the FES current amplitude or pulse width, and the exoskeleton motor torque will be synchronously and simultaneously collected. Session 8: Dynamic walking task with FES or ankle exoskeleton based on the assist-as-needed control approach by combination ultrasound imaging and sEMG. The experimental setup and procedures will be kept the same as mentioned in session 7. The only difference is the old control approach will be replaced by the newly proposed assist-as-needed control approach by combing ultrasound imaging and sEMG signals. The data collection will remain the same as session 7. Among the sessions 5-8 in section 2, after 7 minutes of walking under each speed, a 10 minutes rest period will be provided. Each session may last up to 4 hours, including the motion capture system calibration and devices setup around 45 mins, and walking experiments on the treadmill around 3 hours. The participant may be asked to come back if the data obtained during a session is not found to be satisfactory during analysis. The next visit will be conducted at the participant's convenience and at least after 72 hours. In addition, for safety purposes, for sessions 3, 4, 7, and 8, a safety stop button will be added for the participant to stop the experiment whenever the participant wants to stop the experiments or the participant is to feel uncomfortable during the experiment. For participants with iSCI or transverse myelitis: Before the participant is ready to participate in this study, participants will be asked to fill out a few questionnaires which will collect demographic information and their medical history. The participant will also participate in a physical examination. The physical examination will be completed by Dr. Christine Cleveland, a physician at the outpatient rehabilitation clinic in Chapel Hill. The physical exam and filling of medical history forms will take 60 minutes. The physical examination will consist of a sensory exam (where the participant will be asked to close his/her eyes and report whether if the participant feels a light touch sensation such as a cotton ball and if the participant is able to discriminate a sharp and dull sensation), and a motor exam (where the investigators will examine the participant's ability to move his/her arms and legs). The investigators will also take the participant's heart rate and blood pressure. In case of pregnancy, the participant will become ineligible to participate in the study and will have to discontinue the study. The investigators will screen to check sensitivity or response to FES. The purpose of this test is to allow the participant to experience FES, and based on the experience decide if the stimulation is tolerable and the participant is willing to participate in the study. A test stimulation train will be applied to the participant's lower limb muscle. The participant can stop this test at any time and refuse to participate further. If the participant responds to FES without experience of pain and wishes to participate in this study, the investigators will contact the participant's physician to obtain a medical clearance and verification of the participant's injury level, prior to beginning the actual study. The participant will be required to obtain a signed medical clearance form from his/her physician before participating in the study. Upon receiving clearance from the medical practitioner to participate in this study and if the participant is eligible (please ask Nitin Sharma or the research team member to give the participant an eligibility checklist for Group S) and choose to participate in this study, the participant will be asked to come to the PI's laboratory and BME gait lab to complete the consent form during the first visit.

Voluntary strength, Ankle joint, Neuromuscular model, Ankle exoskeleton, Functional electrical stimulation, Nonlinear control, Sensor fusion, Treadmill walking, Walking biomechanics

The central objective of this study is to validate a new neuromuscular interface that combines surface electromyography and ultrasound imaging to predict human ankle joint voluntary strength or movement intent under isometric and dynamic conditions. The secondary objective is to design control algorithms for a powered ankle exoskeleton and functional electrical stimulation with the consideration of human voluntary strength or movement intent that is estimated from the surface electromyography and ultrasound imaging-based interface. This study is performed with two sets of subjects: people with neurological disorders and people without any neurological disorders. (why have parallel groups)

Individuals with neurological disorders, like iSCI or transverse myelitis, will be recruited (Group S). These individuals usually have weakened ankle joint functionalities but can walk independently.

Surface electromypgraphy-based interface to predict human ankle joint motion intent and use in ankle assistive devices

The study involves the validation of computer algorithms to estimate human ankle joint motion intent and control of ankle joint assistance by using either a powered exoskeleton or an FES system. The ankle joint motions will include seated posture tasks and walking tasks. The instrumented treadmill and Vicon motion capture system will be used to facilitate the cyclic walking pattern and record the participant's kinematics. The human ankle joint volitional effort will be predicted by the sEMG signals from shank muscles. The powered exoskeleton or FES system will provide ankle joint assistance based on an assist-as-needed strategy.

Ultrasound imaging-based interface to predict human ankle joint motion intent and use in ankle assistive devices

The study involves the validation of computer algorithms to estimate human ankle joint motion intent and control of ankle joint assistance by using either a powered exoskeleton or an FES system. The ankle joint motions will include seated posture tasks and walking tasks. The instrumented treadmill and Vicon motion capture system will be used to facilitate the cyclic walking pattern and record the participant's kinematics. The human ankle joint volitional effort will be predicted by the ultrasound imaging signals from shank muscles. The powered exoskeleton or FES system will provide ankle joint assistance based on an assist-as-needed strategy.

Surface electromypgraphy and ultrasound imaging-based interface to predict human ankle joint motion intent and use in ankle assistive devices

The study involves the validation of computer algorithms to estimate human ankle joint motion intent and control of ankle joint assistance by using either a powered exoskeleton or an FES system. The ankle joint motions will include seated posture tasks and walking tasks. The instrumented treadmill and Vicon motion capture system will be used to facilitate the cyclic walking pattern and record the participant's kinematics. The human ankle joint volitional effort will be predicted by combining sEMG and ultrasound imaging signals from shank muscles. The powered exoskeleton or FES system will provide ankle joint assistance based on an assist-as-needed strategy.

The investigators calculate benchmark human volitional effort (torque [N-m]) using inverse dynamics. The investigators predict human volitional effort (torque [N-m]) using neuromuscular model and aforementioned outcome measures - sEMG, ultrasound imaging.

The investigators measure the human ankle position [rad] and velocity [rad/sec] and the desired position [rad] and velocity [rad/sec] using a commercial sensor encoder when the controller is applied.

The investigators measure joint angular position [rad] and velocity [rad/sec] using motion capture system.

The investigators calculate ground reaction forces [N] using load cells installed commercial treadmill.

The investigators utilize surface electromyography (sEMG [V, Hz]) to measure muscle activation level.

The investigators use ultrasound machine to capture muscle's contractility through these parameters: pennation angle of the muscle fibers [deg], muscle thickness [m], and fascicle length [m]. These parameters are further used to predict volitional torque [N-m].

Inclusion Criteria for participants without neurological disorders: Age between the ages of 18 and 64, Weight less than 220 lb, Able to perform ankle movements such as ankle up motion, ankle down motion, side motion towards inside, and side motion towards outside while seated, and Able to walk normally at a preferred speed without any assistive device. Exclusion Criteria for participants without neurological disorders: Any difficulty or an orthopedic condition that would impede ankle movements such as ankle up motion, ankle down motion, side motion towards inside, and side motion towards outside, Any difficulty walking normally or without assistance, Absence of sensation in lower extremities, An allergy to adhesive skin tapes and/or US gels, Pregnant Females, No ankle muscle response to FES. Inclusion Criteria for participants with neurological disorders: 18-64 years of age and have a primary diagnosis of traumatic/non-traumatic iSCI or demyelinating diseases like transverse myelitis, Weight less than 220 lb, Sub-acute or chronic phase (at least 3 months after injury) incomplete motor lesion (AIS C or D at enrollment) at cervical, thoracic or lumbar level, Ability to ambulate over ground independent using either a cane or rolling walker, as well as those that do not require any assistive devices but do have some mobility difficulties, Medically stable with medical clearance for participation, no evidence of cardiopulmonary or pulmonary disease, severe spasticity, and asymmetric hip positions, Ability to respond to FES on dorsiflexors and plantarflexors, and No use of any FES devices or already in use of a FES device for mobility support (like a Bioness device) but will not use the device during the study. Exclusion Criteria for participants with neurological disorders: Subjects with other neuromuscular diseases such as polio, stroke, or multiple sclerosis, Presence of transmissible diseases such as (but not limited to) hepatitis or immunodeficiency virus, Any clinical condition contraindicating gait, Untreatable chronic pain, Severe spasticity (Ashworth scale score > 3), Severe reduction in lower limb joint Range of Motion (ROM) higher than 20 deg, At a high risk of a fracture from osteoporosis, Any skin problem inhibiting robot usage, major depression or psychosis, Subjects with heart conditions and pacemakers, Concurrent severe medical disease, pressure sores, open wounds, existing infection, unstable spine, unhealed limber pelvic fractures, history of recurrent fractures, known orthopedic injury to lower extremities, and osteoporosis, Have open wounds, Pregnant Females, No ankle muscle response to FES.

IPD will not be shared outside of this research group. However, selected data may be published in academic journals, conference papers, or other publications. This data will be de-identified, and will not include the full set of data.

Zhang Q, Kim K, Sharma N. Prediction of Ankle Dorsiflexion Moment by Combined Ultrasound Sonography and Electromyography. IEEE Trans Neural Syst Rehabil Eng. 2020 Jan;28(1):318-327. doi: 10.1109/TNSRE.2019.2953588. Epub 2019 Nov 14.

Zhang Q, Iyer A, Kim K, Sharma N. Evaluation of Non-Invasive Ankle Joint Effort Prediction Methods for Use in Neurorehabilitation Using Electromyography and Ultrasound Imaging. IEEE Trans Biomed Eng. 2021 Mar;68(3):1044-1055. doi: 10.1109/TBME.2020.3014861. Epub 2021 Feb 18.

Zhang Q, Iyer A, Sun Z, Kim K, Sharma N. A Dual-Modal Approach Using Electromyography and Sonomyography Improves Prediction of Dynamic Ankle Movement: A Case Study. IEEE Trans Neural Syst Rehabil Eng. 2021;29:1944-1954. doi: 10.1109/TNSRE.2021.3106900. Epub 2021 Sep 27.

Zhang Q, Sheng Z, Moore-Clingenpeel F, Kim K, Sharma N. Ankle Dorsiflexion Strength Monitoring by Combining Sonomyography and Electromyography. IEEE Int Conf Rehabil Robot. 2019 Jun;2019:240-245. doi: 10.1109/ICORR.2019.8779530.