Using sensory feedback to improve prosthetics
Friday, 28 July, 2023
Listen to the (opens in a new window)podcast.
Sigrid Dupan (above) is a biomedical engineer at the School of Electrical and Electronic Engineering at University College Dublin. She got Science Foundation Ireland and Irish Research Council funding for her project, ‘Examining sensory feedback for upper limb prosthetics’.
It’s more common for people with lower limb loss to use prosthetics than it is for people with upper limb loss.
“There’s a variety of reasons for that,” says Sigrid Dupan. “Partly it's because if you still have one arm, a lot of people can actually do the majority of the tasks they need to do. And prosthetics at the moment are still not that good, so a lot of people choose not to wear a prosthesis.”
Dupan says that “if someone feels good without one, then that’s fine” because “a lot of the time there’s more of a societal pressure to have an arm”. But many more people want prosthetics that enable a full spectrum of movement from “heavy work in the garden to very fine work, which at the moment prosthetics really can’t do”. The demands on future prosthetics are as varied as the people who require them, making it “difficult to find one solution that will work for everyone”.
Dupan elaborates: “For a lot of people, the place they have most trouble is the kitchen where you have a lot of tasks that you need both your hands for. You need to keep something stable and then you need to be able to cut.”
Three types of upper limb prosthetics currently exist: purely cosmetic ones that don’t move; body-powered prosthetics that can open and close the hand via controlling a cable that runs over the person’s shoulder. Thirdly, myoelectric prosthesis - “the ones that you see on the news” - which are controlled by sensors in the prosthetic socket reading muscle activity on the residual limb.
“Unfortunately, they still don't work very well and it's very difficult to control them. People call them brain-controlled, but they're really not. I think there's a disconnect between what people think they can do and what they can actually do.”
Dupan’s work looks at bridging the gap between the reality and the promise of myoelectric prostheses. This research involves using “a lot of coding” to interpret muscle information and experimenting in the lab with the help of people who were either born without a limb or lost one through amputation.
“Someone comes in, they get electrodes on their arm muscles, they sit in front of a screen or there's a prosthetic hand on the table. They will move cursors on the screen or a prosthetic hand with their muscle activity. And then we teach them how to move their muscles to control the prosthesis,” she says. “We also use feedback to make them aware of what they're doing with their muscles because we don't think about what muscles we use for movement. So it takes a while for people to be aware of what they're actually doing with their muscles. That's a big part of what we do. It takes a lot of programming beforehand and a lot of analysis afterwards to see how we can best teach people to do this.”
People using myoelectric prosthetics need to learn how to use their muscles in “a more efficient way”, which Dupan likens to taking up a new sport.
“If you learn tennis, you won’t be very good at it the first day. But if someone really explains to you how to hold your racquet and what to do, you will improve. It's the same thing. It's a new way of moving muscles and of controlling the prosthetic.”
Understanding sensory feedback plays a crucial role in the process.
“If you are aware of what you're doing right or wrong, then you can train using the prosthesis and over days and weeks people get better and better at it.”
Training the prosthesis is a machine learning technique where only the prosthesis learns new things, not the user, she clarifies.
Dupan has worked with people who want prosthetics to be “more comfortable and lighter” and also “more robust and accurate”. Even those who choose not to wear a prosthesis now are interested in future advancements.
For instance, current myoelectric prostheses cannot yet judge grip force.
“So people wearing them need to always track everything they do with their eyes because they can't feel how hard they're pressing. If they are holding a glass it might start slipping and they need to pay attention so they can press harder immediately. Otherwise the glass falls.”
Dupan speculates that following the war in Ukraine there may be more demand for prostheses.
“A lot of advancements in this field started in the US and were very much related to wars. They have a lot of veterans who need prosthetics and so they put much more money into the development of new prosthetics. It's a part of our field and work that feels bitter; that wars have to happen for money to be available.”
In the coming decades, Sigrid would like to see a “much, much broader range” of prosthetics available to chime with the wide variety of people’s wants and needs.
“I think a lot of the time, the changes in prosthetics - and I'm definitely also part of this - come from engineers sitting together going, 'how can we most closely get to a human hand?' But maybe that's not what people need? There will be a need for hands that closely resemble human hands for some people. And finer control of those would be amazing. But if we can broaden what we do in prosthetics and how we envision prosthetics I think that will be very important.”
Listen to the (opens in a new window)podcast.
This article was originally published on 25 October 2022.