Upper-Limb Prosthetics: Seeking the Sense of Touch
October 2013 Issue
I feel, therefore I am.
Although philosopher and mathematician Reneì Descartes actually said, "I think, therefore I am," feeling, not only in the sense of emotions but also in the literal fifth sense, touch, helps us interact with the world around us and integrate our sense of self-of who we are as complete beings. As researchers and developers continue in their quest to meld human and machine into a single functional entity, restoring a sense of touch opens the door to creating more intuitive, easier to use, and more efficient prosthetic limbs. An enormous benefit is that users would not need the intense visual concentration currently required to operate a myoelectric prosthesis because touch perception would enable them to know that they were grasping the object securely but not so tightly as to crush or break it.
Research shows that augmenting the prosthesis with sensory perception helps users to integrate the device into their "sense of self " as a natural part or extension of their own bodies rather than as an attached, alien object or tool, thus increasing acceptance of the prosthesis and promoting a positive body image. These are important factors, considering the historically high rejection rates for upper-limb prostheses, ranging from a bit less than 30 percent to as high as an estimated 50 percent.
"Tactile feedback is essential to intuitive control of a prosthetic limb, and it is now clear that the sense of body self-identification is also linked to cutaneous touch," Paul Marasco, PhD, a researcher and sensory neurophysiologist at the U.S. Department of Veterans Affairs (VA) Advanced Platform Technology (APT) Center, Cleveland, Ohio, and his colleagues point out in the article, "Robotic Touch Shifts Perception of Embodiment to a Prosthesis in Targeted Reinnervation Amputees" (Brain. 2011 March; 134 (3):747-58).
The authors investigated the rubber hand illusion, first used by researchers in 1998. As the subject looks at the rubber hand, the researcher strokes the subject's own hidden hand. The subject begins to perceive the location of his or her hand shifting toward the rubber hand. Marasco and colleagues tested two individuals with amputations who had undergone targeted muscle reinnervation (TMR) surgery, using a robotic touch interface coupled with a prosthesis. "Measurements provide evidence that the illusion [of the prosthesis perceived as their natural hand] created is vivid.... [T]his may help amputees to more effectively incorporate an artificial limb into their self image, providing the possibility that a prosthesis becomes not only a tool, but also an integrated body part."
TMR Surgery Leads to Serendipitous Discovery
A hotbed of upper-limb prosthetics research is bubbling around the world as scientists and developers work to add touch sensations to prosthetic hands.
One promising initiative resulted from an unexpected side effect of TMR surgery at the Rehabilitation Institute of Chicago (RIC), Illinois. The TMR procedure involves transferring residual-arm nerves to alternative muscle sites. After the nerves successfully reinnervate the new site, these muscles can produce electromyogram (EMG) signals to control a prosthesis by thought. In 2006, the technique provided the surgical platform that enabled Jesse Sullivan, who has bilateral upper-limb amputations, to become the first "bionic man" to use a mind-controlled prosthesis successfully.
Then another exciting development occurred unexpectedly. A few months post-surgery, as an assistant wiped Sullivan's chest with alcohol, Sullivan felt the sensation of cold-on his missing hand. "The nerves we were using to innervate the muscles were mixed nerves-a combination of efferent fibers sending motor signals to the muscles and afferent fibers receiving sensory information," explains Jon Sensinger, PhD, director of RIC's Prosthesis Design & Control Laboratory. "We were surprised that the afferent fibers also reinnervated the skin."
With the opportunity to add sensory input to a prosthesis, the RIC team developed targeted sensory reinnervation (TSR). Patients can perceive varying temperatures, sharpness of objects, vibrations, and pressures on their reinnervated skin as though these stimuli were occurring on their missing limb. "We were excited about this because it gave us a direct sensory portal to the hand."
The team evaluated portal stability and quality of sensations. "Even several years post-TSR surgery, patients still had the same, consistent sensory portal," Sensinger says. The quality of the signal was lower in the hand but higher in the arm, "which makes sense because we only had the sensory receptors that were in the arm, not the high density of receptors in the hand, but we did have the added power the brain's cortex uses to process hand sensations."
To convey sensation, the team used tactors, which did an excellent job of delivering various sensations gathered by sensors in the socket, such as pressure, temperature, and force and delivered to the reinnervated skin and thence to the brain, Sensinger says. However, much to the team's surprise, patients didn't find the sensory input particularly helpful. "So we took a step back at that point."
The researchers have recently taken up the sensory input challenge again. The team says it is excited about electrical stimulation research at Case Western Reserve University (CWRU), Cleveland, Ohio. "There's a group led by Dustin Tyler that is developing some really innovative protocols," Sensinger comments, explaining that Tyler's research is producing sensation that actually feels realistic rather than the tingling sensation usually produced by electrical stimulation. "It's a promising approach, and we're looking forward to collaborating with them in providing the sensory portal to the missing hand."
Another angle is determining which type of sensory feedback is most important to prosthetic users. "We found that we cannot use a variety of sensory feedback in our real-time manipulation of objects because it's simply too slow," Sensinger explains. "What we now believe is that the brain uses sensory feedback to form models of how the world works." The brain uses inverse dynamic models to learn, for instance, how to ride a bicycle. Once the skill is learned and the model is created, people can ride a bicycle easily. "If feedback is removed, models can't be generated as easily."
Using computational motor control, a multidisciplinary research approach to discover the principles of human motor control, researchers have achieved considerable success in describing how able-bodied persons interact with their world, Sensinger points out. As research continues, he says, "We're hopeful that we will be able to describe how persons with amputation generate those models and what type of sensory feedback, such as pressure or force, would be most useful to them."
DARPA Marches Forward
Under the aegis of the Defense Advanced Research Projects Agency (DARPA) Reliable Neural-Interface Technology (RE-NET) program, innovative projects in advanced prosthetic limbs have used both brain interfaces and muscle and peripheral nerve interfaces. Currently, brain interface research is limited to persons with paraplegia, but peripheral nerve interface technology may soon be available to individuals with amputation for advanced prosthetic control.
"The novel peripheral interfaces developed under RE-NET are approaching the level of control demonstrated by cortical interfaces and have better biotic and abiotic performance and reliability," says Jack Judy, PhD, DARPA program manager, as quoted in a DARPA press release May 30. "Because implanting them is a lower risk and less invasive procedure, peripheral interfaces offer greater potential than penetrating cortical electrodes for near-term treatment of amputees. RE-NET program advances are already being made available to injured warfighters in clinical settings." Among other initiatives, research at RIC, CWRU, and Northwestern University, Chicago, has been included under the RE-NET program.
Stimulating Real-Life Touch
The gradations in the natural hand's ability to perceive touch sensations involved in feeling textures, vibrations, movements, temperatures, pain, discomfort, pressure, and 3D shapes verge on the infinite. However, the sensation generally experienced in current sensory-input prostheses has been described more as a tingle, buzz, or poke-still a long way from the natural hand.
The research group led by Dustin Tyler, PhD, associate professor, Department of Biomedical Engineering, CWRU, mentioned by Sensinger, is taking a new approach to electrical stimulation by fine-tuning the information cutaneous sensory nerves perceive and transmit to the brain to create a natural sense of touch as the prosthesis is used to manipulate objects. Besides enabling a much more natural sense of touch, the CWRU team's efforts have enabled amputee subjects to control a prosthesis successfully without intense visual concentration on its movements. Test subjects have even been able to successfully pick up and move objects without being able to see them.
The research team is using flat interface nerve electrodes (FINEs) which flatten out nerve fibers so that several of them can be exposed at once to electrical currents providing feedback, explains Jeff Blagdon in a May 31 article in The Verge, an online technology and innovation magazine (www.theverge.com). The electrodes are implanted around the nerves rather than within them, thus preventing nerve damage.
"The location around the nerves determines where [the subject] feels it; how we stimulate determines what is felt," Tyler explains. "One way can make him feel as if someone is laying a finger across his hand; with another, he'll feel a vibration as though he were running his fingers across the teeth of a comb. Other stimulation has felt like different surface textures such as sandpaper or Velcro®. So he has a variety of sensations depending on how we do the stimulation." Pressure sensors also are positioned on the prosthetic hand so that the amputee subject feels when he grabs an object and how tightly he is holding it. He has also been able to distinguish between stimulated sensations and the phantom sensations he experiences, Tyler notes.
To test the effect of sensation without visual cues, subjects have successfully completed the Box and Blocks Test for manual dexterity in which users pick up blocks from one side of a box and place them in the other, while their vision is occluded and white noise blocks out prosthesis motor auditory cues. "Without the sensation, he just moves the prosthesis and closes it as though he were picking up a block and moving it," Tyler says. "When we turn the sensation on, he knows whether or not he has that block." The research team has also employed other tests, such as pulling one grape off a bunch of grapes. "You don't want to squeeze or crush the grape, yet you need to hold it tightly enough to be able to pull it off the stem."
What Tyler would like to do soon is partner with some manufacturers to make a prosthetic hand with the sensors built in. Once that is available, users can take the electric stimulator home and use it on a daily basis rather than in short laboratory sessions. "We want to show that results are repeatable and that it's generally cost-effective and workable for upper-limb amputees.
Then we want to move into a complete system that can be integrated with their prosthesis and is on all the time," Tyler says.
SynTouch: Reflexive Touch Control
Possibly close to beginning commercialization is a product from SynTouch, Los Angeles, California, a startup company begun in 2008 by researchers from the Medical Device Development Facility of the University of Southern California, Los Angeles. The company's flagship product, the BioTac®, is a tactile sensor system consisting of a rigid core surrounded by an elastic skin filled with a fluid to give a compliance "remarkably similar to the human fingertip," according to the company. The sensors, electronic circuitry, and connections are protected inside the core; no sensors are placed in or on the skin itself.
SynTouch researchers tested a commercially available myoelectric prosthetic hand modified to include BioTac® sensors. Compared to the subject's usual myoelectric prosthesis, the BioTac contact-detection method resulted in faster completion times to grasp and move a set of fragile objects.
SynTouch will present a paper detailing research results during November's 2013 International Conference on Intelligent Robots and Systems (IROS) in Tokyo, Japan.
The BioTac sensors are rather expensive, sophisticated devices primarily used for research and robotic applications, points out SynTouch CEO and founding partner Gerald E. Loeb, MD. For prosthetic application, the team found that most of the benefit was achieved by incorporating a control that was largely subconscious, setting off reflexive responses in grasping objects. The biologically inspired algorithm allowed the user to grasp fragile objects securely by generating a simple EMG signal.
When the subject was fitted with tactors, "he actually sensed all the sensory inputs from the BioTac, but felt they were too distracting," says Matt Borzage, SynTouch founding partner and head of business development. "What he just really wanted to do was to be able to grasp different objects easily with his prosthesis." With just the reflexes the subject was able, without concentrated visual attention, to grasp a variety of fragile objects without breaking, crushing, or dropping them, according to Borzage.
Working from these results, SynTouch has developed a simpler, reduced-function version, the NumaTac, which it exhibited during the May 2013 Institute of Electrical and Electronics Engineers (IEEE) International Conference on Robotics and Automation (ICRA) in Karlsruhe, Germany. "The contact-detection algorithm can be integrated into an existing prosthetic hand system simply by intervening in the signal that comes from the myoelectric electrode detectors and modulating that system as it goes into the controller," Loeb says. "We don't have to make any changes to the hardware and software, which is very helpful because of the regulatory requirements involved in medical device engineering."
For these and other new prosthetic technologies to enter the clinical mainstream and benefit patients, they must navigate the complex and often lengthy road to commercialization. Prosthetic component manufacturers keep an eye on these developments and, for promising projects, can help shepherd technologies into the marketplace. Ottobock, headquartered in Duderstadt, Germany, for example, "has many collaborations worldwide and is always evaluating new technologies that have the potential to provide benefits to our patients," says Kevin Kelley, international project coordinator for Ottobock Healthcare Products, Vienna, Austria. "It is a long and difficult road from innovative concepts to commercial production of medical devices, and so we often work with innovators to guide this process. At the end we go through a tough and thorough evaluation of the resulting technology, looking at technical, competitive, legal, and financial aspects in order to determine whether it makes sense to incorporate [it] into a commercial product." As research and development teams from many different disciplines push back the boundaries of prosthetics in the quest to replicate the natural hand, individuals with upper-limb amputations and O&P clinicians can expect exciting developments to come.
Miki Fairley is a freelance writer based in southwest Colorado. She can be contacted via e-mail at