Sections

Weather Forecast

Close

ALS patients banking on technology to help them keep speaking

The voice banking computer program displays sentences for clients to read as well as the quality of each recording. Steve Kuchera / Forum News Service1 / 2
Karen Stubenvoll, who has ALS, records a sentence during a recent voice banking session at UMD’s Department of Communication Sciences and Disorders. People banking their voices record 1,600 sentences. Steve Kuchera / Forum News Service2 / 2

DULUTH, Minn. - Karen Stubenvoll is at sentence 1,214.

"I know that's real fun," she reads into a headset in a clear, steady voice, and then pauses to clear her throat.

Seated next to her in a soundproof booth in the old Chester Park School, Jolene Hyppa Martin makes a subtle hand signal. Stubenvoll resumes.

"The Velveteen Rabbit had no hind legs at all."

Another pause.

"This was to be an eventful day."

Stubenvoll, a retired doctor, is well on her way toward the 1,600 sentences she ultimately will have read to develop a synthetic version of her own voice for the time when she may need it.

Stubenvoll, 59, was diagnosed a year ago with amyotrophic lateral sclerosis, or ALS, a progressive disorder affecting the nerves and muscles. It afflicts about 20,000 Americans at any given time, according to the ALS Association.

After she was diagnosed, Stubenvoll retired as a hospitalist for Essentia Health and resigned as chairwoman of the board of Hawk Ridge Bird Observatory. She remains an enthusiast of the nature site and still manages some hiking, she said, using trekking poles to support herself. She has had to give up bicycling, and traveling has become much more difficult.

She hasn't experienced another symptom that sometimes accompanies ALS. That is dysarthria, a slurring of speech. Stubenvoll still speaks with perfect diction in a voice that seems slightly hoarse after reading sentences for two hours, but easily understandable.

But speaking can become tiring, and Stubenvoll can foresee a time when it could become worse.

"I've noticed my voice getting a little weaker recently," she said. "When I say my voice is getting weak, I think a lot of it is because my breathing is getting weaker. So by the end of the day I'm just more tired when I'm talking, and I'm starting to feel a little hoarse."

She also uses a machine to help her breathe 11 hours each night and an hour or two each afternoon, and she can't speak when she's using it.

When she learned the Robert F. Pierce Speech-Language-Hearing Clinic at the University of Minnesota Duluth offers what's known as voice banking, she couldn't see a downside, Stubenvoll said. Since she lives in Duluth, she could easily get to the clinic. And the ALS Association currently is underwriting the cost.

How it works

Stubenvoll is the fifth client to use voice banking at the clinic, which is in the old elementary school on the UMD campus. Hyppa Martin, an assistant professor in UMD's Department of Communications Sciences and Disorders, began offering the service last spring.

It works like this:

Over a number of sessions, Stubenvoll reads the 1,600 stock sentences into the microphone. She said she recognizes the sources of a number of those sentences — "The Velveteen Rabbit," "The Wizard of Oz," "White Fang," "Little Women." After that's done, she will add some customized information, such as the pronunciation of her last name.

When it's all finished, Hyppa Martin will send the recording to the Nemours Speech Research Laboratory at the Alfred I. duPont Hospital for Children in Delaware, the developer of what it calls a ModelTalker System.

In a few days, a rough version of a synthetic voice for Stubenvoll will come back to the clinic. She and Hyppa Martin will listen to it and decide on the refinements they wish to make, and then send it back to Delaware. A final version will come back, and Stubenvoll's synthetic voice will be loaded into an iPad.

"It's not like recordings where you can play a recording back," Hyppa Martin said. "It's an actual text to speech that will sound like a computerized version of me. Or a computerized version of Karen."

The process, from the time the recordings are completed, typically takes less than two weeks, Hyppa Martin said.

Once she has the synthetic version of her voice, Stubenvoll will be able to type what she wants to say if it's difficult or impossible for her to talk. Instead of a Siri-like voice, people who know her will be able to hear her familiar, warm tones.

"Your voice is a really personal thing," Hyppa Martin said. "People recognize you by your voice, and when you make a phone call people would, before the days of caller ID, probably recognize you by your voice. So a lot of people want to retain their voice and retain their individuality."

The voice can be put on any sort of device, she said. It can be installed in technology that tracks eye movement to allow for hands-free computer operation. Entire messages — prayers, humorous sayings and the like — can be recorded in what's known as message banking.

Frequently used words can be stored along with typically used phrases. For example, Stubenvoll could say "Let's go to Hawk Ridge" at the touch of a button. She can have "phrase boards" for certain situations, such as eating out or talking on the phone with family.

Clients "have the ability to say whatever they want to say right when they want to say it," Hyppa Martin said.

Other uses

There also are ways to use the technology for people who have never had their own voice, she said. For example, if a child has a developmental disability that prevents her from speaking, a friend or family member of about the same age and from the same region could "donate" her voice.

Hyppa Martin is doing that herself.

"Last weekend I was in the Cities and met with my friend who has cerebral palsy, and I'm donating my voice to her," she said.

Hyppa Martin already has completed banking her own voice for that project.

Hyppa Martin learned voice banking from her colleagues at the Twin Cities campus and is the only person at the UMD clinic currently certified to do it. In the Twin Cities, students are trained in the process and earn clinical hours by working on voice banks with clients. But in Duluth, there aren't enough clients yet to ensure the students would get the hours they need, Hyppa Martin said.

Among the five clients so far, one came from the western edge of the Iron Range and another from north of the Iron Range, she said. Aware of the geographic challenge, Hyppa Martin hopes to take voice banking to people who would have to make long drives to the clinic, and she is working with two undergraduate students on an experiment to see if that's feasible. Using the sound shield from a drum set, egg crate foam and "some clamps you get at Menard's," they've engineered a traveling sound booth, Hyppa Martin said. They're in the process of comparing results from the portable sound booth with results from the clinic's permanent sound booth.

"We wouldn't want to start banking somebody's voice in their home and then have the quality not be sufficient," she said.

Based on population numbers, the ALS Association anticipated voice banking in Duluth might attract one participant per semester, Hyppa Martin said. With five people served in less than a year, the clinic is ahead of that projection.

"It's been substantially more popular than we had anticipated, but we can still help more people than we're serving right now," she said.

To learn more

To find out more about voice banking or to set up an appointment, call the Robert F. Pierce Speech-Language-Hearing Clinic at (218) 726-8199 or email cd@d.umn.edu.

Advertisement