Lecture 17: Advanced Neuromuscular Interfaces for Control of Prosthetic Limbs Flashcards
5 different signals can be recorded with a permanently implanted device in animal. What is best to record reliably over decades?
* ENG
* EMG
* Force transducer
* length change devices
* single unit electrode
ENG (electroneurogram); a nerve cuff implanted correctly will last decades
EMG: intramuscular electrode will eventually break down as muscle shape is constantly change, especially with large forces.
Force transducer: signals decline over weeks and a few months. new tendon fibres grow around the transducer and eventually force was being carried by new tendon fibres and not the ones the transducers were wrapped around.
length change devices are good for a few months before they break. they are attached to the bone with bone screws and include a column of liquid within the device which will change with motion allowing a nice signal.
single unit electrodes (motor unit): are ok for a few days or weeks only, they’ll move
Why in a neurogram are you primarily recording afferent nerve traffic?
Although neurograms record from mixed nerves (like femoral) afferents dominate visible traffic patterns because they have a low firing threshold and thus are constantly reaching threshold for firing.
small and large motor neurons have higher firing threshold and thus aren’t activated nearly as much, only when recruited.
In the amputee case study, why did researchers implant a nerve cuff around the ulnar nerve, distal to last functional motor branch?
What did they find? What did they do after?
- test feasibility and practicality of graded stimulation of amputated nerve to provide sensory feedback of grip strength from the electrical hand
- test feasibiity of using motor signals generated in ulnar nerve for neuroelectric control of powered prosthesis. Could patient selectively activate forearm muscles if he thought of them (flexor carpi ulnaris, pronator teres, supinator, extensor carpi radialis)?
Findings
* could activate desired muscles
* when researchers stimulated ulnar nerve, indviidual reported appropriate sensation. reported feeling sensation in ring and small fingers. he could feel 7 levels of stimulation
Also, Implanted bipolar EMG electrodes over the bellies of flexor carpi ulnaris, extensor carpi radialis, pronator teres, and supinator. All signals ran through a percutaneous connector
Resulted in individual being able to cause 2 movements at once (ex: opening/closing hand & wrist rotation). allowed for greater control and smoother movements
Why did the farmer’s arms fail? What did they have to do? What ended up happening
Developed an infection? after 13 months due to the percutaneous connector weeping.
paper said skin surrounding percutaneous sensor broke down
observed skin retraction, thinning, and necrosis around the connector
friends and family would hit his arm, damaging it and causing connector to weep?
Wires had to be cut and the implanted device was left in in hopes that they could implant a radiocontrol after receiving clearance. Never received clearance but was eventually fitted with another prosthetic (not as good0
Discuss how conventional low and high frequency electrical stimulation can treat phantom limb pain
- what is the benefit of high frequency over low frequency?
Low frequency
* delivery of electrical stimulation to the spinal cord or targeted peripheral nerve to block the sensation of pain (Gate Control Theory)
* stimulation of Aβ fiber causes activation of inhibitory interneuron that supresses ascending pain pathway.
* 40–60 Hz range, and produces the sensation of paresthesia, which must overlap the area of pain to provide analgesia
High frequency
* stimulation pattern selectively activates inhibitory interneurons in the dorsal horn (DH) of the spinal cord without stimulating the AB fibers responsible for paresthesia.
* * supresses hyperexcitable wide dynamic range neurons (WDR), which are sensitized and hyperactive in chronic pain states.
* 10kHz stimulation at low amplitude (1-5mA), and short duration/pulse width (30microsec)
- Paresthesia-free pain. paresthesia can be uncomfortable or intolerable to some patients. can allow for patient to drive (since you’re not allowed to drive with paresthesia-based stimulation
Pain pathway:
Normally (in absence of pain), inhibitory interneuron is tonically activated, causing supression of the pain pathway. During strong pain, c fibres are activated strongly and inhibits inhibitory interneuron, allowing signals to be sent up the ascending pain pathway. Ab are activated by non-noxious stimuli, such as light touch, pressure, and hair movement
Describe Daniel Tan’s study with neural cuffs, sensations, and amputees following study
Created a neuronal interface in prosthetic (20 channels total):
* 4 contact cuff on radial nerve
* 8 contact fine cuff on median and ulnar nerve (fine cuff = flattens nerve, increases surface area you can put electrodes on)
demonstrate that a simple electronic cuff placed around nerves in the upper arm of two amputees wearing a replacement neuroprosthetic limb directly activates the neural pathways responsible for hand sensations. Could feel different modalities and in the appropriate areas
sensory feedback also eliminated pain in the phantom hand
To record afferent signals, should you use ENG or EMG?
Why is it hard to record motor signals in amputated nerves?
To record afferent signals, nerve cuff electrodes (ENG) are the most stable (last for several decades).
Cut axons that terminate in a neuroma survive, but remain markedly atrophied, generating extremely small action potentials that may not be differentiable from background EMG and noise. But, if you graft the cut axons proximal to the neuroma to a new muscle, the axon diameter will increase again upon forming new functional connections.
If you record the EMG signals from these cross-reinnervated muscles, signals will be amplified and recordable.
Describe the case study for a woman with a proximal amputation.
* what was the surgical procedure
* what was the result?
Denervate pectoralis major and minor in preparation for reinnervation. Strip away fat under the skin overlying the chest musculature to improve recorded EMG signals.
Grafted musculocutaneous nerve, median, ulnar, and radial onto large chest musculature like pectoralis major (2 heads), and minor.
Allowed nerve to sprout and reinnervate target muscle area
After allowing some time for the nerve to establish new connections, place EMG electrodes on surface of skin of these muscles and record signals to see if person could generate voluntary signals when just thinking about moving the arm, wrist, shoulder.
When asked to think of 4 different movements (elbow flexion, extension, hand closing, hand opening) impulses were detected by surface EMG recordings within the prosthetic harness. These signals were forwarded to the computer processor in the prosthetic to make the arm perform movement. Now she can control the elbow, wrist, and shoulder synergistically!
Some months later, found out that when they touched her chest, she reported it felt like they were touching her pinky. Removal of fat allowed afferents to sprout and innervate percutaneously. Sensation in the hand was re-routed to skin on chest, giving patient a sense of touch.
How long after grafting would it take for reinnervation? Several weeks to a couple months, depending on distance.
Describe the case study (T5): transhumeral amputation
Dennervate medial bicep and lateral tricep.
Median nerve transferred to medial bicep. Distal radial nerve transferred to lateral tricep
Summarize the following study: Pre-emptive Treatment of Phantom and Residual Limb Pain with Targeted Muscle Reinnervation at the Time of Major Limb Amputation.
What were the recommendations of this study? Why is this not practical?
Objective: determine whether targeted muscle reinnervation at the time of major limb amputation decreases the incidence and severity of phantom limb pain and residual limb pain.
Study design: 51 patients underwent intervention compared to 438 controls. Primary outcomes included an 11 point NRS and Patient-Reported Outcomes Measurement Information
System (PROMIS) pain intensity, behaviour, and interference.
Result: Patients who underwent TMR had decreased the incidence and severity of phantom limb pain and residual limb pain
Preemptive surgical intervention of amputated nerves with targeted muscle reinnervation should be considered to reduce PLP and residiual limb pain.
Difficult to do though because it requires extensive pre-planning, large surgical team (neurologist, orthopedics, etc.), and longer time under anesthesia. If patient prevents emergently or late and limb needs to be removed immediately, may not have time to pre-plan TMR.
Describe the following study: regenerative peripheral nerve interface allows realtime control of artificial hand in upper limb amputee
Design: divided large peripheral nerves into several fascicles and grafted them separately onto different denervated muscles in the stump. Axons increased in diameter. This created a regenerative peripheral nerve interface to bioamplify efferent motor action potentials.
also implanted electrodes at the same time to record inside the stump (as opposed to recording EMG on the surface of skin), epimyseal electrodes sutured over the nerve.
3 month process of regeneration, revascularization, and reinnervation
Describe the following study:
An osseointegrated human-machine gateway for long-term sensory feedback and motor control of artificial limbs
Rationale: conventional socket suspension for high level amputations can be uncomfortable, heavy, and limit range of motion. You also require a harness to hold the prosthetic in place.
Osseointegrated human-machine gateway allows for unrestricted limb motion.
Drill a hole in the end of cut bone and place osseointegrated fixture.
Prosthetic limb has an abutment screw which will attach to the osseointegrated fixture. Also put in epimyseal electrodes on biceps, brachialis, and triceps. Placed nerve cuff on ulnar nerve. Connector also implanted in upper arm.
Result: improved range of motion and prosthetic control
Describe the study on neural feedback strategies to improve grasping coordination in neuromusculoskeletal prostheses
developed a prosthetic that provided tactile perception during routine grasping and grasping under uncertainty.
sensors in prosthetic that can “feel” grip and load force. load force is a measure of slip. controls grip.
load force/slip sensors check for microslips which will trigger reflexes to increase grip quick before it falls out of ur hand.
fingertip sensors measures force at thumb, index, and middle finger and modulate grip
Sensory feedback that can modulate grip creates a close loop system allowing for greater control
Researchers tried several different algorithmic models but the one for best grip was one that did not modulate grip until sensory signal was detected. no anticipation is better
Describe how a prosthetic can be used to restore sensation
Sensors in tip of bionic finger detect and code infromation of texture in the form of current stimulus that can be fed back to the user via implanted transversal interfasicular multichannel electrodes. The signal can also be communicated to user via flat interface nerve electrodes.
Can code for both rough and smooth surfaces
One study tied the remnant tendons of the quadricep and hamstring muscle groups together in above the knee amputees. Why was this done? How could this be applied for below the knee amputations?
creates a agonist-antagonist myoneural interface
when one group contracts it, it automatically stretches the antagonist group, generating proprioceptive afferent signals in the antagonist group. This will provide the CNS with sensory input to improve motor control. Prosthetic feedback is communicated to peripheral nervous system through FES of the antagonist muscle to control position or force applied on mechanically linked agonist muscle
Below the knee amputation
Create an agonist-antagonist myoneural interface by attaching the tendons of peroneus longus and tibialis posterior to each other. This will help with control of the subtalar joint (inversion/eversion). For ankle joint control (plantarflexion/dorsiflexion), attach lateral gastrocnemius to tibialis anterior
allow person to travel more controlled on uneven terrain
