Research Projects
The Neural Engineering Center for Artificial Limbs has many projects underway. The focus of our research is to improve the control of artificial limbs as well as to understand the neurophysiological mechanisms of prosthesis use.
-
-
-
-
-
-
-
Overview of Accomplishments to Date with Targeted Reinnervation
Clinical Application of Targeted Reinnervation in Shoulder Disarticulation Amputees
Targeted reinnervation surgery has been performed successfully on shoulder disarticulation (SD) amputees and is now a clinically available option for them. In SD patients, targeted reinnervation involves the connection of the residual brachial plexus nerves (which carry the neural information to and from the arm*) to the pectoral muscles of the chest. Depending on the anatomy, additional muscles such as the serratus anterior muscle and the latissimus muscle may also be used. It takes several months for the nerves and muscles to form functional connections. Eventually, neural commands intended for the missing limb are successfully routed to the new target muscles. This results in contractions of the target muscles when the patient attempts to move his/her missing limb (view
Targeted Reinnervation Effect on Chest Muscle Movement).
*The musculocutaneous, median, radial and ulnar nerves of the brachial plexus are used for reinnervation. The musculocutaneous nerve normally innervates elbow flexors such as the biceps muscle. The median nerve normally innervates wrist, palm and finger flexors, wrist pronator muscles, and intrinsic hand muscles. The radial nerve normally innervates elbow extensor muscles, such as the triceps muscle, as well as the wrist and finger extensors of the forearm. Finally, the ulnar nerve innervates wrist and finger flexors as well as many intrinsic hand muscles. Results from ulnar nerve reinnervation have been inconsistent.
Clinical Application of Targeted Reinnervation in Transhumeral Amputees 
Targeted reinnervation has also been performed successfully on transhumeral amputees and is a clinically available option for them as well. For transhumeral amputees, the surgery involves changing the innervation of two muscles of the upper arm, the short head of the biceps and either the brachialis or the lateral head of the triceps. The original innervation of these target muscles is cut and the median and distal radial nerves* are connected. The innervation of the lateral biceps (elbow flexion) by the musculocutaneous nerve and the remaining triceps (elbow extension) by the proximal radial nerve are left intact. Due to the limited number of available muscle sites, the ulnar nerve has not yet been reconnected during targeted reinnervation on transhumeral amputees.
* see note above.
Prosthesis Control
The electrical signals generated by contractions of the reinnervated muscles (or a combination of reinnervated and native muscles) can be recorded using surface electrodes and used to control an artificial limb. The prosthesis is initially controlled using conventional or “direct” control. Here, one electrode pair placed over each muscle is responsible for controlling a single motion of the prosthesis (e.g. elbow flexion or hand open/close). The surgery generally results in four independent myoelectric controls sites which correspond directly to prosthesis movements and are therefore intuitively controlled by the patient (prior to targeted reinnervation surgery, only one to two myoelectric control sites are available and in the case of shoulder disarticulation amputees, these sites are physiologically unrelated to the prosthesis movements). A more complex control scheme using additional electrodes and pattern recognition algorithms allows for the control of additional and more complex prosthesis movements, such as hand grasping patterns (see Advanced Signal Processing Methods to Enhance Prosthetic Limb Control with Targeted Reinnervation ).
Functional Outcomes
Several measures are used to assess the performance of targeted reinnervation subjects following recovery and training with their new prosthesis (see
Clinical Training and Outcome Measures for Upper Limb Prostheses ). Subjects show improved performance with their new control scheme. For instance, subjects have shown two and three-fold improvements on the box and blocks test as well as increased speed on the clothespin relocation test. AMPS testing on one subject revealed an increase from 0.3 to 1.98 for motor skills and 0.9 to 1.98 for process skills. Patients report that they are able to perform tasks previously impossible for them prior to reinnervation surgery. For a bilateral shoulder disarticulation subject who underwent targeted reinnervation and subsequent fitting and testing, these tasks included taking out the garbage, putting on a hat, vacuuming and shaving.
Related Publications
Miller LA, Stubblefield KA, Lipschutz RD, Lock BA and Kuiken TA. Improved myoelectric prosthesis control using targeted reinnervation surgery: a case series, IEEE Trans Neural Sys Rehab Eng. , 16(1):46-50, Feb 2008.
Kuiken TA, Miller LA, Lipschutz R, Lock B, Stubblefield K, Marasco P, Zhou P and Dumanian G. Targeted reinnervation for enhanced prosthetic arm function in a woman with a proximal amputation, Lancet , 369(9558):371-80, Feb 2007.
Hijjawi JB, Kuiken TA, Lipschutz, RD, Miller LA, Stubblefield KA and Dumanian GA. Improved myoelectric prosthesis control accomplished using multiple nerve transfers, Plast Reconstr Surg., 118(7):1573-78, Dec 2006.
Lipschutz RD, Kuiken TA, Dumanian GA, Miller LA Dumanian GA, and Stubblefield KA. Shoulder disarticulation externally powered prosthetic fitting following targeted muscle reinnervation for improved myoelectric prosthesis control, J Prosthet Orthot., 18(2):28-34, Apr 2006.
Kuiken TA Dumanian GA, Lipschutz RD, Miller LA and Stubblefield KA. The use of targeted muscle reinnervation for improved myoelectric prosthesis control in a bilateral shoulder disarticulation amputee, Prosthet Orthot Int., 28(3):245-253, December 2004.
Clinical Application of Targeted Reinnervation in Transradial Amputees
We are just beginning to extend targeted reinnervation to transradial amputees. Because transradial amputation is the most common type, this application could potentially benefit thousands of amputees. Targeted reinnervation in a transradial amputee would be a simple surgical procedure to transfer the residual nerves from the intrinsic hand muscles to remaining forearm muscles. This has the promise of giving the amputee intuitive control over the prosthetic wrist and hand, with the potential to control multiple degrees of wrist movement and more complicated grasping patterns.
Related Publications
O’Shaughnessy KD, Dumanian GA, Lipschutz RD, Miller LA, Stubblefield KA and Kuiken TA. Targeted reinnervation to improve prosthesis control in transhumeral amputees. A report of three cases . J Bone Joint Surg Am., 90(2):393-400, Feb 2008.
Advanced Signal Processing Methods to Enhance Prosthetic Limb Control with Targeted Reinnervation
Deciphering Neural Control Information
A major task in improving prosthetic limb control with targeted reinnervation is to determine how much information can be extracted from the reinnervated muscles. High-density electrogmyogram (EMG) signals are collected with >100 electrodes placed over reinnervated muscles while subjects attempt different movements. The amplitude maps of these recordings demonstrate clear changes in muscle activation patterns with different intended movements.
Electrocardiogram (ECG) signals contaminate the EMG signals of shoulder disarticulation targeted reinnervation subjects; therefore, improved techniques are needed to remove them. Adaptive filters and clipping filters have been developed and successfully implemented to remove ECG contamination in real-time.
Pattern classification methods are applied on high-density EMG recordings to match the muscle activation patterns with the intended movements. Over 90% classification accuracy can be obtained in deciphering between 16 different intended upper-limb movements, including thumb and finger movements.
Channel Reduction and Real-Time Control of Prostheses
Although high-density EMG recordings provide the maximum amount of control information available from the reinnervated muscles, this technique is not clinically feasible. Channel reduction algorithms can be used to reduce the number of electrodes needed to preserve the necessary neural control information. We have found that sufficient information for accurate classification can be extracted by 8 to 12 strategically placed electrodes.
A real-time pattern recognition-based prosthesis control system has been developed in collaboration with the Institute of Biomedical Engineering at the University of New Brunswick (UNB) (opens new window) which uses up to 16 optimally placed EMG electrodes as inputs. The intended movement of the user can be decided up to every 30 ms using an algorithm which classifies the signals recorded on the EMG channels. The speed of the motor which controls the corresponding joint on the prosthesis is proportionally controlled by the magnitude estimation involving all EMG signals. We are currently working to improve the classification algorithms and make them adaptable and more robust to slight changes in electrode location, environmental conditions, and changes in muscle contraction patterns over time.
A 2-dimensional virtual reality system created at UNB includes an arm with a three degree of freedom (DOF) shoulder, one DOF elbow, three DOF wrist, and hand capable of seven different grasping patterns. The system is used to test real-time control algorithms. This is an efficient way to assess the functionality and usability of each new design. It also allows subjects to practice real-time control without a prosthesis.
Real-time control of prostheses with targeted reinnervation is also tested using multifunctional prosthetic arms. We collaborate in the testing of state-of-the-art prosthetic arms developed outside of the RIC. This contributes to the future design of prosthetic arms, the development of control schemes for use with targeted reinnervation, and the continued care of our patients.
Related Publications
Huang H, Zhou P, Li G and Kuiken TA. An analysis of EMG electrode configurations for targeted muscle reinnervation based neural machine interface , IEEE Trans Neural Sys Rehab Eng., 16(1):37-45, Feb 2008.
Zhou P, Lowery M, Englehart K, Huang H, Li G, Hargrove L, Dewald J and Kuiken TA. Decoding a new neural-machine interface for control of artificial limbs , J Neurophysiol., 97(5), Nov 2007.
Zhou P, Lock B and Kuiken TA. Real time ECG artifact removal for myoelectric prosthesis control, Physiol Meas., 28(4):397-413, Apr 2007.
Zhou P, Kuiken TA. Eliminating cardiac contamination from myoelectric control signals developed by targeted muscle reinnervation , Physiol Meas., 27(12):1311-27, Dec 2006.
Quantification of Targeted Sensory Reinnervation
After targeted reinnervation surgery, patients slowly develop new sensations in the skin overlying their reinnervated muscles. After several months, stimulation of this skin reliably results in the perception that their missing limb is being touched. We have termed this phenomenon "transfer sensation," and are investigating it in order to better understand the mechanisms of sensory reinnervation as well as to define design requirements for prosthetic feedback devices.
Each patient who has developed transfer sensation is tested repeatedly in an effort to map the referred sensation on his or her reinnervated skin as well as to find the pressure threshold at which these referred sensations are first perceived.
Subjects are also tested on their reinnervated skin to determine the just-noticeable difference in forces they can perceive at different baseline levels, their ability to perceive the difference between a sharp or blunt stimulus, the threshold at which they can first perceive high and low frequency vibrations, their ability to detect hot and cold stimuli, and the smallest electrical current at which they first experience transfer sensation.
Related Publications
Kuiken TA, Marasco PD, Lock B, Harden RN and Dewald JPA. Redirection of cutaneous sensation from the hand to the chest skin of human amputees with targeted reinnervation, PNAS, 104(50):20061-20066, Dec 2007
Neural Plasticity after Targeted Reinnervation
Studies are underway in collaboration with
Dr. Julius Dewald and the
NeuroImaging and Motor Control Lab (opens new window) to map the brain activity of amputees before and after targeted reinnervation surgery. The effect of re-routing the sensory and motor nerves of the hand on the cortical representations of these areas is not yet understood. This paradigm offers a new opportunity for investigating cortical plasticity in adults.
Clinical Training and Outcome Measures for Upper Limb Prostheses
Clinical Training
Following recovery from targeted reinnervation surgery, patients are fit with a prosthesis which takes advantage of the new myoelectric control sites. Initial training is essential for fine-tuning electrode placements and settings in the new prosthesis. Training is important for insuring that patients take full advantage of the benefits offered by the nerve-transfer surgery. Specifically, subjects are taught to use the direct relationship between the movements they command and the movements of the prosthesis. Patients learn to operate multiple degrees of freedom simultaneously and to integrate the use of their prosthesis with their able arm by performing two-handed activities. The goals of training are to increase the ease and fluidity of prosthesis use as well as to decrease the mental load required of the user. This is ultimately expected to increasing the wearing time and usefulness of the prosthesis.
Outcome Measures
Patients are routinely tested during visits to the NECAL lab during the months and years subsequent to targeted reinnervation surgery. The goals of testing are to quantify the functional benefits offered by the procedure as well as to chart the patient’s functional improvements over time. Subjects also help us assess the performance of new and experimental arms and new control paradigms. Standard measures include performance on the box-and-blocks and clothespin relocation tests. New clinical measures are continuously being assessed for possible inclusion in the standard testing procedure.
Development of Neural Interfaces for Lower Limb Prostheses
We are currently in the process of evaluating the potential application of targeted reinnervation to the control of lower limb prostheses. The targeted reinnervation procedure would provide access to the neural commands intended for the distal muscles of the leg lost by amputation. These command signals would be used in conjunction with feedback from the powered leg to give the amputee more accurate and intuitive control. Pattern recognition could be used to robustly detect a change in the user’s intended movement, e.g. from level walking to going up stairs.