Therapy Interventions for Upper Limb Amputees Undergoing Selective Nerve Transfers

Summary

This work presents a protocol to enhance prosthetic function after selective nerve transfer surgery. Rehabilitation interventions include patient information and selection, support of wound healing, cortical re-activation of sensory-motor areas of the upper limb, training of selective muscle activation, prosthetic handling in daily life, and regular follow-up assessments.

Abstract

Targeted Muscle Reinnervation (TMR) improves the biological control interface for myoelectric prostheses after above-elbow amputation. Selective activation of muscle units is made possible by surgically re-routing nerves, yielding a high number of independent myoelectric control signals. However, this intervention requires careful patient selection and specific rehabilitation therapy. Here a rehabilitation protocol is presented for high-level upper limb amputees undergoing TMR, based on an expert Delphi study. Interventions before surgery include detailed patient assessment and general measures for pain control, muscle endurance and strength, balance, and range of motion of the remaining joints. After surgery, additional therapeutic interventions focus on edema control and scar treatment and the selective activation of cortical areas responsible for upper limb control. Following successful reinnervation of target muscles, surface electromyographic (sEMG) biofeedback is used to train the activation of the novel muscular units. Later on, a table-top prosthesis may provide the first experience of prosthetic control. After fitting the actual prosthesis, training includes repetitive drills without objects, object manipulation, and finally, activities of daily living. Ultimately, regular patient appointments and functional assessments allow tracking prosthetic function and enabling early interventions if malfunctioning.

Introduction

High amputations of the upper limb provide a challenge for prosthetic replacement¹. Aside from elbow joint function, active prosthetic systems should include opening/closing of the prosthetic hand and ideally also pronation/supination and/or wrist extension/flexion. However, the control of standard myoelectric devices usually relies on the input signals from two muscles only². These are traditionally the biceps and triceps muscles after transhumeral amputations and the latissimus dorsi and pectoralis major muscles after glenohumeral amputations³. To control all prosthetic joints, amputees need to switch between the active joints (e.g., by using a co-contraction of the two muscles)¹. While this provides a stable control paradigm, a significant restriction ensues with resulting slow and unintuitive control, which does not allow simultaneous movements of two or more prosthetic joints⁴. This limits the functionality of the prosthesis and is one of the reasons for high prosthetic abandonment rates after above-elbow amputations⁵.

To overcome limited and unintuitive control for these types of prosthetic fittings, selective nerve transfers can be utilized. This approach, also known as Targeted Muscle Reinnervation (TMR), consists in surgically establishing myo-control signals by re-routing nerves that initially served the amputated hand and arm to different target muscles within the residual limb^6,7. After successful reinnervation, more selective activation of the reinnervated muscle units becomes possible⁸. The resulting electromyographic (EMG) activity can then be used for prosthetic control and can yield up to six control signals.

While there is a broad agreement that TMR can significantly improve prosthetic function⁹, selective activation and appropriate control of multiple muscles in the stump pose a challenge to patients, especially in the early post-operative period. This enhanced complexity of prosthetic control paired with the reduced multisensory feedback following amputation requires a specific rehabilitation to fully benefit from the surgical procedure. Here, a step-by-step guideline for the therapy interventions is provided based on recent recommendations¹⁰. An overview of the interventions and the estimated time they take in an ideal setting can be found in Figure 1.

Figure 1: Overview of stages within the rehabilitation process, including the milestones that mark the start of a new stage. Please click here to view a larger version of this figure.

Protocol

The protocol was developed within a European Delphi study¹⁰. The assessment of its application on patients was approved by the local research ethics committee of the Medical University of Vienna and carried out according to the Declaration of Helsinki. If not mentioned otherwise, the steps described here should be carried out by an occupational therapist or a physiotherapist.

1. Pre-surgical interventions

1. See the patient for a multidisciplinary consultation.
   NOTE: The medical core team should include a surgeon, an occupational therapist and/or physiotherapist, a prosthetist, and a psychologist.
2. Collect the patient history (reason and date of amputation, previous medical/therapeutic interventions after amputation, co-morbidities, general medical history, prosthetic satisfaction) and ask about expectations for prosthetic rehabilitation and demands for a prosthetic system in daily life.
3. Check for relevant inclusion and exclusion criteria.
   1. Consider the patient for TMR if they fulfill the following criteria: amputation above the elbow, good general health, personal desire for good prosthetic function, willingness to participate in post-surgical therapy for up to 15 months.
   2. Exclude patients who have any untreated psychological co-morbidities.
4. Perform a physical examination of the residual limb, focusing on skin and soft tissue problems, neuromas, range of motion, and possible additional nerve injuries.
   NOTE: If surgical interventions for the residual limb are needed (e.g., soft tissue corrections), the surgeon addresses them during TMR surgery.
5. Evaluate the patient's general fitness regarding whether they will be able to carry a myoelectric prosthesis after TMR (~3 kg) and determine further interventions they may need during rehabilitation (such as strengthening the limb, exercises for endurance, or trunk stability). If the patient has a prosthesis, assess its function, preferably with standardized assessment instruments.
6. Evaluate the patient's mental well-being and recognize psychiatric diseases, such as depression or post-traumatic stress disorder (psychologist). If the evaluation shows the need for treatment, ensure as a team that the patient receives it.
7. Based on the patient's needs, patient history, and examination, discuss available prosthetic options with the patient. Ensure that the patient understands that TMR involves a lengthy rehabilitation, where active involvement is needed.
8. Determine whether TMR is the best option for the patient. Provide the patient sufficient time to consider different options and/or discuss them with friends and family.
9. See the patient again (either in the full interdisciplinary team or as a rehabilitation professional with the surgeon) to plan the procedure unless the patient had already decided on TMR during the initial consultation.
10. If the medical team and the patient agree that TMR should be performed, ensure that financial reimbursement of the whole process is guaranteed and that rehabilitation and prosthetic fitting will be organized.
11. See the patient for therapy sessions before surgery. According to the patient's need, include exercises for pain treatment, endurance, body symmetry, trunk stability, strengthening of the limb and posture, and motor imagery tasks.
    1. Additionally, train one-handed activities and support the patient with assistive devices, which can be helpful to support independence in activities of daily living.
       NOTE: At least one pre-surgical therapy session is recommended. More may be needed to address specific problems. If only a short time before surgery is available, include particular interventions in the post-surgical therapy.
12. Perform the TMR surgery (surgeon)⁹.

2. Early post-surgical interventions

1. In the first days after surgery, mobilize the patient and ensure that they regain their physical fitness. Once more, remind the patient that reinnervation might take ~ 3-6 months.
   NOTE: Early post-surgical interventions should happen once or twice daily while the patient is hospitalized. If the patient can perform the below-listed interventions by themselves after initial explanations, a therapy session once a month is sufficient and can also happen as teletherapy in an online setting. Otherwise, seeing the patient twice a week for 30-60 min is recommended.
2. Treat possible edemas in the area of the surgery by bandaging, using custom-made liners, propping up the residual limb and/or offering lymphatic drainage. Ensure that the patient receives adequate pain medication.
3. Start with scar treatment (scar cream application and scar massage) when the wounds are fully closed. Improve the range of motion in the shoulder joint for transhumeral amputations by passively moving the arm and instructing the patient to perform active exercises using the full range of movement.
   NOTE: The patients are asked to use the scar cream available to them; nothing specific is recommended.
   1. Communicate all interventions with the surgeon and/or see the patient with the surgeon at least once.
4. If the patient had a prosthetic fitting before surgery, evaluate whether it can be re-fitted. If needed, have a prosthetist change the socket or replace the electrodes in a myoelectric fitting.
   NOTE: In some cases, a re-fitting of the socket might not be feasible.
5. Facilitate the reinnervation process on a cortical level with methods such as mirror therapy^11,12, imagined movements¹³, or lateralization training¹⁴ (or follow the structure of Graded Motor Imagery, which includes these interventions¹⁵) to activate the sensory-motor cortex areas responsible for the upper limb.
   NOTE: This enables the patient to activate reinnervated muscles more efficiently at a later point in therapy.
   1. For mirror therapy, set up a mirror in front of the patient and ask them to hide the residual limb behind the mirror. Instruct them to perform different movements with the healthy hand while watching its reflection in the mirror.
   2. Ask the patient to imagine different movements of the amputated hand and arm while keeping their eyes closed. If helpful, ensure that the patient can do this in a quiet, undisturbed environment.
   3. For lateralization training, present the patient with cards that show either left or right hands and arms. Ask the patient to name the side and give the patient feedback on their choice.
      NOTE: If the patient prefers novel technologies, introduce the patient to computer programs or apps that provide the same function.
6. Continue any pre-surgical interventions depending on the patient's needs.

3. Signal training

1. Study the surgery report to understand which muscle parts are reinnervated and which nerves were transferred. Understand that the transferred nerve determines the motion(s) the patient needs to attempt to activate the reinnervated muscles (e.g., any muscle innervated by the ulnar nerve can be activated by imaging hand closing or wrist flexion after successful reinnervation).
   NOTE: If anything is unclear, meet the surgeon to discuss the nerve transfers and the rehabilitation plan.
2. Three months after surgery, start testing for the first volitional contractions of the reinnervated muscles. If an activity can be found, continue with the steps below, and aim to see the patient for weekly or bi-weekly therapy sessions until surface EMG (sEMG) control is mastered. If no activity can be found, continue with the early post-surgical interventions, and perform another test a few weeks later.
   1. For evaluating volitional muscle activity, set up a system for sEMG biofeedback.
      NOTE: Here, a system that can display up to six EMG signals and allow an individual amplification of each channel is preferred.
   2. Prepare the patient's skin to reduce impedance by removing excessive body hair, dead skin flakes, oil, or skin cream¹⁶. Explain the goal of the assessment and the functionality of the system to the patient.
      NOTE: Plan therapy sessions for 30 min or less at this stage. Otherwise, muscles may become easily fatigued, and the patient may lose needed focus. If short sessions are not possible, mix different therapy interventions (EMG and posture training) to avoid fatigue. Figure 2 displays a standard setup for EMG biofeedback training.

Figure 2: Setup for surface EMG biofeedback. The therapist places an electrode on the patient's skin where the EMG signal is expected while explaining the needed movement cue (making a fist). The patient and therapist can see the patient's muscular activity (EMG) on the computer screen and use this feedback for finding the best electrode position and movement cue. Please click here to view a larger version of this figure.

3. Instruct the patient to perform hand and arm movements depending on the original function of the donor's nerves (e.g., hand closing if the ulnar nerve was used) and try to palpate the muscle.
4. Place a surface EMG electrode on the skin above the muscle. Consider reinnervation to be successful if the signal amplitude during activation is 2-3 times higher than during relaxation¹⁷.
5. If such activation is not possible, instruct other movements connected to the donor's nerve (e.g., flexion of the wrist or pinkie finger, if the ulnar nerve is the donor) and move the electrode slightly above the recipient's muscle.
6. Repeat the evaluation for volitional activation with all nerves based on the surgery reports and note which muscles can be activated and with which motor command. Ask the patient to train the motor commands at home.

3. Train the selective activation of the reinnervated muscles.
   1. Use EMG biofeedback to display the activity of one muscle. Ask the patient to think of the previously evaluated movement patterns and use an sEMG electrode (see Table of Materials) to pick up the recipient's muscle signals.
   2. Use the notes from the previous evaluation. If easier for the patient, ask them to perform the desired movements bilaterally.
   3. As soon as the patient can repeatably activate the muscle, train muscle relaxation as well.
      NOTE: Muscle relaxation corresponds to EMG amplitudes close to zero and is sometimes hard to achieve.
   4. Ask the patient to activate the muscle and fully relax it repeatedly. Make sure that there are 5-10 s of break in between the activations.
   5. Instruct the patient to perform different movements and vary electrode positions to find the combination leading to the highest amplitude (hotspot). Take a photo of the best position or mark it on the skin.
   6. If more muscles can already be activated, train the activation and relaxation of each muscle individually.
   7. After a reasonable control of the single muscles is possible, display the activity of two muscles. Start with antagonistic muscles/movements such as hand opening and closing. Instruct the patient to activate one muscle while the other one should be as relaxed as possible.
   8. Try different movement cues for both muscles if such a selective activation is not possible. Explain to the patient that selectivity needs some training and make enough time for this step.
   9. As soon as the selective activation of two muscles is achieved, add a third muscle and repeat the previous steps. In the same way, add one muscle at a time until the patient can selectively activate each one. Plan several therapy sessions to train this.
      NOTE: To allow direct simultaneous prosthetic control at a later stage, the patient needs the ability to repeatedly activate each muscle while maintaining no/very little activation of all others. Figure 3 shows a schematic drawing of the excellent separation of six different signals in an EMG biofeedback system.

Figure 3: Schematic drawing of the EMG signals displayed via biofeedback. Every channel (with a different color) is mapped to a specific muscle part and will later be responsible for a particular prosthetic movement. Good separation, as depicted here, ensures that the prosthesis only performs intended movements. Please click here to view a larger version of this figure.

4. Once the selective activation of all signals is established, introduce a table-top prosthesis as shown in Figure 4.
   NOTE: Some systems allow to display EMG signals while moving the prosthesis simultaneously. These systems are preferred for training as they enable more precise feedback.

Figure 4: Patient controlling a table-top prosthesis with surface electrodes mounted on his residual limb. Please click here to view a larger version of this figure.

1. First, only enable one prosthetic joint, e.g., the hand, and ask the patient to control it while carefully watching the prosthesis. If the prosthetic hardware allows it, explain to the patient that a low EMG amplitude corresponds to slow movement while fast movement is achieved through a high signal. Let them test different movement speeds.
2. Change the active prosthetic joint (e.g., elbow joint or wrist) and let the patient control these levels with their EMG signals.
3. Once a good control of the single levels is possible, switch on all prosthetic joints and enable simultaneous control. Instruct the patients that unwanted prosthetic movements are normal at this initial stage of prosthetic control.
   NOTE: A light activation of their muscles may support selective control of single prosthetic joints.
4. When this is mastered, give the patient a first impression of grasping with a prosthetic device (the table-top prosthesis) by holding objects (small balls, tubes of bottles) close to the open prosthetic hand and asking them to close.
5. If they want, let the patient play with grasping and releasing objects they hold with their unaffected hand (for unilateral amputations). Let the patient know that sometimes failing to grasp or release objects is normal but should improve with training.

5. Ensure that a certified prosthetist provides a test-fitting with all electrodes for myoelectric control placed in the socket correctly.
   1. To support the correct electrode placement in the socket, mark the EMG hotspots on the patient's skin and note the prosthetic movements for each hotspot.
   2. If possible, see the patient with the prosthetist for the plaster casting and answer any questions the prosthetist might have regarding electrode placement.
   3. When the first (test-) socket is ready, check its fit together with the prosthetist. Ask the patient to wear it and report any issues with the fitting (such as too much pressure at specific points). Check the electrode positions by connecting the electrodes in the socket to an EMG biofeedback system or a table-top prosthesis and asking the patient to control it.
   4. If no sufficient control of the table-top prosthesis is possible when wearing the socket while it can be done with electrodes mounted on the skin, re-assess the electrode positions in the socket with the prosthetist and change them (and/or the socket) if needed.

4. Prosthetic training

1. Once the (test-)socket fits well and the patient can control a table-top prosthesis with the electrodes embedded in the socket, ask the prosthetist to assemble the complete prosthetic fitting.
2. See the patient with their new prosthetic fitting together with the prosthetist and surgeon. Check the fit of the prosthesis, discuss with the team whether the changes are needed, and answer any questions the patient might have.
   1. Explain the basic functionality of the prosthesis to the patient, such as degrees of freedom, how switching between the active joints works (if needed). Also, explain whether the prosthesis is waterproof and how it should be cleaned.
3. Train donning and doffing of the prosthesis.
   NOTE: The duration and frequency of prosthetic training depends on the complexity of the prosthetic fitting, the therapist's experience, and the patient's motor learning ability. Needed changes in the socket (e.g., for electrode positions) may delay the training. In optimal settings, the patient attends therapy twice a week for 30-60 min in the first few weeks and has the option of using the test fitting for home training in between.
4. Train prosthetic movements without external objects.
   1. Ask the patient to perform easy movements of the prosthesis, such as opening/closing of the hand. If possible, connect the prosthesis via Bluetooth to its software to display EMG signals.
      NOTE: If the prosthesis does not react to the patient's motor commands or performs unintended movements, use EMG biofeedback to figure out the reason for this. If the problem is hardware-related (socket-fit or electrode placement), contact the prosthetist to solve this. Otherwise, try adapting the software settings and/or instruct the patient to adjust their motor commands (e.g., slighter contraction).
   2. Continue with training single movements of all prosthetic joints as described in step 3. If the prosthesis allows different movement speeds, instruct the patient to vary the speed of movement. Ensure that the patient is precisely doing what they intend to do.
   3. To add more complexity, ask the patient to control the prosthesis in different positions (standing, sitting, or with varying shoulder positions for transhumeral amputees) and combine more degrees of freedom simultaneously (e.g., closing the hand flexing the elbow at the same time).
5. Train object manipulation
   1. Provide the patient with different objects such as stress balls or wooden blocks. Explain that the manipulation of objects adds another layer of complexity.
      NOTE: Normally, the patient needs to train for a while to have complete control over the prosthesis while working with external objects.
   2. Ask the patient to use their healthy hand (for unilateral amputees) to put the object into the prosthetic hand. Then instruct to close the prosthetic hand, move the prosthetic elbow and/or wrist joint and, finally to release the object.
   3. As a next step, place the objects on the table/shelf/etc. Ask the patient to pick them up with the prosthetic hand and place them somewhere else.
   4. Finally, tasks that require more precision, such as stacking wooden blocks or grasping a ball rolling on a table, can be trained.
6. Train activities of daily living
   1. Ask the patient which common activities (such as carrying a bag, doing the laundry, cooking, dressing, eating with cutlery, opening/closing a door, etc.) they regularly do in their daily life. Prioritize a few of them and train them in therapy.
      NOTE: Discuss that the prosthesis cannot be used for bathing and showering.
   2. For training daily activities, suggest performing them with the prosthesis based on the experience (e.g., with some prosthetic hands, it is easier to pick up small objects if the hand is in a maximum pronated position). Let the patient perform the tasks based on suggestions provided. If they have other ideas on how to perform them, let the patient try their approach and encourage them to try many strategies and be creative.
      NOTE: It is essential to explain to patients that prosthesis training takes time and patience.
   3. Give the patient feedback on the performance during task completion. The feedback should be based on compensatory movements (little to none is preferred) and the patient's time to perform the task. If you or the patient are dissatisfied with how the task could be completed, try different strategies.
   4. Ask the patient which further, more specific activities are essential in their daily lives (e.g., sports, leisure activities, childcare or specific tasks required for their jobs) and discuss how they can use the prosthesis within these tasks.
      NOTE: If possible, directly train a few of these tasks with the patient during the therapy sessions (either in the clinic or at the patient's home environment). Not all tasks can be performed with a prosthesis. In some cases, specific prosthetic fittings or assistive devices are needed (e.g., for some sports or playing instruments). Although there were significant advances in recent years, prosthetic devices are still far from equivalent to human hands in function¹⁸.
   5. Ask the patient to use the prosthesis at home and make notes (or photos and videos) of tasks they are doing or they feel they cannot do.
   6. Use these notes to discuss different strategies for prosthetic use in the following therapy sessions.
   7. Repeat prosthetic training within the therapy sessions and at home until the therapist and the patient understand that the prosthesis can be used well in daily life.
   8. Discharge the patient from therapy.

5. Follow-up assessments

1. Invite the patient to a multidisciplinary medical consultation at 3 months after discharge from rehabilitation.
   1. Ask the patient how they use their prosthesis at home and work and discuss any problems.
   2. If the patient reports any problems, discuss/provide solutions for them.
2. Assess the patient's prosthetic function by using standardized tests (such as the Southampton Hand Assessment Procedure (SHAP)¹⁹, the Action Research Arm Test (ARAT)^20,21, or the Assessment of Capacity for Myoelectric Control (ACMC)^22,23). Ask the patient to fill out standardized questionnaires for quality of life and hand use in daily life (such as Short Form 36 (SF-36)²⁴ and Disabilities of Arm, Shoulder and Hand (DASH)²⁵).
3. If the test results show a problem, discuss this with the patient and offer solutions for their problems (if possible).
4. After the first follow-up consultation, invite the patient every 6 months to a multidisciplinary consultation and structured assessments to ensure ongoing good prosthetic function.

Representative Results

The described rehabilitation protocol was implemented in a clinical setting at the Medical University of Vienna, and its feasibility and outcomes were assessed in a clinical study, which was recently published⁹. As reported⁹, 30 patients participated in the trial to evaluate the feasibility of TMR surgery and subsequent rehabilitation. Figure 5 displays that out of these 30 patients, 11 underwent TMR as a pain treatment rather than a means to improve function via prosthetic fitting. Out of the remaining 19 patients originally aiming for a prosthetic fitting, five decided against it due to the high costs of the fitting (estimated between 75,000-150,000 €), insufficient time for rehabilitation, or high weight of the prosthesis. In one patient, intra-operative exploration revealed a global brachial plexus injury, making further nerve transfers impossible. This patient kept using his body-powered device. Of the remaining 13 patients undergoing prosthetic rehabilitation, 10 were available for a follow-up assessment.

Figure 5: Flowchart showing the patients included in the feasibility study. Please click here to view a larger version of this figure.

Outcomes were assessed using the Southampton Hand Assessment Procedure (SHAP)¹⁹, the Action Research Arm Test (ARAT)^20,21, and the Clothespin-Relocation Test (CPRT)^6,26. These assessments are commonly used tests to evaluate prosthetic function. The evaluation took place at least 6 months after the final prosthetic fitting. Additionally, patients were asked about their prosthetic wearing habits.

As described by Salminger et al.⁹, assessment of the 10 patients after TMR surgery revealed a SHAP score of 40.5 ± 8.1 (with a healthy upper extremity having a score of about 100) and ARAT score of 20.4 ± 1.9 (with 57 being the maximum score and 0 representing no upper extremity function) (Table 1). In the CPRT, patients were able to complete the tasks within 34.3 ± 14.4 s. They reported wearing their prosthesis daily with a wearing time ranging from 3-10 hours per day.

Outcome assesment	Score	Expected score for healthy upper extremity
SHAP	40.5 ± 8.1	100
ARAT	20.4 ± 1.9	57
CPRT	34.3 ± 14.4 s	-

Table 1: Prosthetic function of patients following TMR surgery and rehabilitation. In the SHAP and ARAT, higher scores mean a better function, which is also indicated by less time needed in the CPRT. Total patient assessed: n = 10. Adapted with permission from Reference⁹.

Discussion

In recent years, selective nerve transfers have been increasingly used to enhance prosthetic function²⁷. Experienced clinicians in this field have come to appreciate that rehabilitation is essential to enable amputees to use a prosthesis after the surgical procedure skilfully²⁷. However, there is a lack of structured therapy programs. The current protocol aimed to provide the occupational and physical therapists with the tools and structure to guide the patients throughout the long TMR process. In contrast to previous suggestions for therapy (developed for less complex nerve transfers)²⁸, there is a stronger focus on pre-prosthetic training and the use of EMG biofeedback to allow selective muscular control.

As shown in the feasibility study⁹, discussing the patient's expectations is essential for post-operative success. The inclusion of highly motivated patients certainly helped to achieve the described excellent outcomes. Less compliance to the described protocol might result in reduced prosthetic function. Additionally, not all patients wish to receive a prosthetic fitting (or can afford to get one). However, TMR may still be feasible to improve neuroma or phantom limb pain since recent studies have shown the potential of nerve transfers to alleviate these conditions^29,30,31. For such cases, the rehabilitation program is foreshortened. Still, we have experienced that regular training of controlled activation of the reinnervated muscles and a prosthesis can further improve the pain situation³². Here, shared decision-making is essential as some patients might wear a prosthesis for its potential to reduce pain in the long term³², while others might not be interested.

In our experience, a detailed discussion with the patient is essential to evaluate future compliance. Depending on the reinnervation time, motor learning capacity, and the patient's availability, the rehabilitation process is likely to take between 9-15 months. Suppose a patient does not strive toward the improvement of upper limb function or might make better use of another device (e.g., body-powered prosthetics). In that case, one might not consider the time (and possibly financial) commitment worth it. To save resources, we strongly recommend only including patients who express a strong interest in the procedure and only perform the surgery for functional purposes when the full rehabilitation procedure is anticipated. Finally, the costs for the surgery, therapy, and fitting should likely be covered at that point.

The described study protocol needs to be adapted for each individual based on clinical reasoning to meet their specific needs. Physical and psychological co-morbidities need to be considered and adequate treatment (e.g., psychotherapy) offered in addition to the interventions described here. In patients receiving TMR immediately after amputation, a closer screening for psychological conditions developing overtime may be needed. Apart from this, no change in the protocol is required for this group of patients. They might even progress faster in motor learning as they might still be used to bimanual activities. Within this protocol, the nerve transfers operated by the surgeon define, which motor commands need to be trained and are expected for which muscle parts. The choice of the prosthetic end device influences prosthetic training. For multi-articulated prostheses, switching between different grasp types and how to use them needs to be included in therapy, if necessary.

For patients living far away from the clinical center or those who cannot attend in-person rehabilitation regularly, adoptions in the rehabilitation protocol are needed. They include a stronger focus on home training, the possible involvement of a therapist near the patient's home, and telerehabilitation sessions via online video calls. Solutions for telerehabilitation need to provide a stable video and audio connection while fulfilling all data protection requirements. In these patients, a first visit to the clinical center should be planned at 6-9 months after surgery for signal training. The visit is usually for 1 week, with therapy sessions twice a day. In a majority of cases, good signal separation can be achieved at this time. Otherwise, another stay for signal training is needed, and the patient may get a simple sEMG biofeedback device for home training. When good signal separation is established, the prosthetist can fabricate a test socket, and the signal positions can be defined during the stay. This allows the prosthetist to create the final fitting when the patient returns home. The final prosthesis can be fitted in a second 1-week visit 1-2 months later, and prosthetic training can be initiated. Advanced prosthetic training and further follow-up visits can either happen in a remote setting or during a further visit to the center, depending on the patient's needs.

Furthermore, other surgical interventions, such as osseointegration³³ to improve the mechanical interface for the prosthesis, can be combined with TMR³⁴. If this is the case, specific interventions must be included (such as the graded weight-bearing training after osseointegration³⁵). Additionally, while the described protocol is intended for direct prosthetic control systems (where one electrode corresponds to one movement), its principles remain the same if a pattern recognition control system is planned. The main difference in rehabilitation is that the selective activation of single muscles becomes less relevant, while particular and repeatable activation patterns of several muscles need to be trained³⁶.

Materials

Name	Company	Catalog Number	Comments
Dynamic Arm Plus® system with a Variplus Speed prosthetic hand	Ottobock Healthcare, Duderstadt, Germany		This prosthetic system was used together with a computer (and Bluetooth connection) for sEMG Biofeedback. Later, it was used for table top prosthetic training and as the patient's prosthetic fitting.
ElbowSoft TMR	Ottobock Healthcare, Duderstadt, Germany		In combination with the Dynamic Arm Plus system and a standard computer (with Windows 7, 8 or 10), this software allows the visualisation of EMG signals as well as changing settings in the prosthetic system.
EMG electrodes	Ottobock Healthcare, Duderstadt, Germany	electrodes 13E202 = 50	The EMG electrodes used in this study were bipolar and included a ground and a 50 Hz filter. They were used with the Dynamic Arm Plus®.
Folding Mirror Therapy Box (Arm/Foot/Ankle)	Reflex Pain Management Therapy Store		This box was used for mirror therapy.