A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study

The overall goal of this procedure is to train upper limb amputees to use advanced prosthetic limbs. This is accomplished by using the amputees inherent learning strategies to guide rehabilitation. The first step is to have the patient imitate and repeat movements performed by a demonstrator.

Next, the patient engages with a computer system to reinforce learning and to train algorithms for prosthetic control. The final step is for the patient to independently control an advanced prosthetic limb. Ultimately, structured rehabilitation training is used to show that using amputees inherent learning strategies improves advanced prosthetic control.

So I first had the idea for this method when we are attempting to train an amputee to control an advanced prosthetic limb. Unfortunately, the existing pattern recognition methods were not effective. Visual demonstration of this method is crucial as the method itself is fully visual, and it's important for clinicians as well as engineers to consider it while training patients and not only algorithms.

Begin by fitting a customized socket and prosthesis to the participant. Calibrate the prosthetic hardware and use the collected calibration data for prosthetic control. Do not give the participant any formal training, however, instruct them that they will perform eight actions, which include one resting state with the prosthetic limb that will allow control of a visual target on a computer screen.

Display a static picture of the movement on the screen, and ask the participant to follow a visual cue. Then show the participant the unique and specific polar plots which correspond to their electromyography or EMG patterns. Encourage the participant to follow the visual cue by pointing to the screen.

Repeat the task three times with different arm positions. Relaxed laxed reaching in front and reaching across. To enhance system training, complete all eight actions and three force levels for each arm position to collect a total of 72 individual samples of system training.

Once the training is complete, use the training data sets from this session to calibrate and adjust the prosthesis so the participant may practice realtime control. Allow the participant to control the prosthesis by sequential proportional control or one movement at a time with the speed of the device proportional to the levels of muscular contractions. Once each of the eight actions is performed in a repeatable, reliable manner, switch the control scheme to proportional and simultaneous control, allowing more than one movement of the wrist at a time.

Then have the participant practice simple tasks such as picking up a simple object and laying it on its side. Allow two hours of rest. Before completing the outcome assessment, assess the participant's performance using the Southampton hand assessment procedure or SHA outcome Measure.

Compare SHAP scores to a baseline measure using the participant's standard prosthesis. Ensure that the participant does not have access to the customized prosthesis and control algorithms after the completion of the naive session. Three months after the naive session performed, the structured training session structure the training session in the following order, imitation, repetition, reinforcement, and prosthetic control.

For imitation, instruct the participant to directly imitate the desired eight actions performed by the demonstrator in real time. To do this seat the participant at a 45 degree angle from the demonstrator and provide an unobstructed view of the demonstrator hand matching the affected side of the participant. Execute each action for three seconds.

For repetition. Ask the participant to repeat the action that has been imitated 10 times for 30 seconds. Without any visual cues, observe the corresponding EMG activity as represented by the polar plots of each movement.

Then for reinforcement, present the participant with visual feedback of their eight actions exactly as was seen in the naive session. Ask the participant to perform the actions while viewing the realtime polar plots with recorded motion overlays to reinforce learning. Ensure that the reinforcement training is exactly the same as the training given during the naive session.

Once the training is complete, use the training data sets from this session to calibrate and adjust the prosthesis so the participant may practice real-time control. Allow the participant to freely control the prosthesis by sequential proportional control or one movement at a time with the speed of the device proportional to the levels of muscular contractions. Once each of the eight actions is performed in a repeatable, reliable manner, switch the control scheme to proportional and simultaneous control, allowing more than one movement of the wrist at a time, have the participant practice simple tasks such as picking up a simple object and laying it on its side.

Allow 24 hours of rest before completing the outcome assessment. Finally, assess the participant's performance using the shap outcome measure. Compare SHAP scores to a baseline measure using the participant's standard prosthesis.

This protocol describes a structured rehabilitation method, which includes imitation, repetition, and reinforcement learning strategies for improved multifunctional prosthetic control. Here SHAP scores are compared between the baseline naive session and structured training session. The results of the shap outcome test showed that the greatest improvements were during tasks that used flexion and extension movements of the wrist, which this new control scheme and customized prosthesis provided over the participant's standard prosthesis.

After this procedure, upper limb amputees can quickly learn to control simultaneous movements of an advanced prosthetic limb while maintaining the safety and security of traditional proportional control across all movements.