A NEW FREEDOM OF SPEECH
App helps give ALS patients their voice back
ALS patients who lose their voices can soon have them back in a new way thanks to a partnership between Google and Cambridge-based ALS Therapy Development Institute that uses speech recognition technology to restore the independence robbed by the disease.
“My wife and sons typically need to translate for me if nonfamily are in the conversation. Telephone voice conversations are not possible. Social isolation seems inevitable,” wrote ALS patient Mark Seger, who is an active participant in ALS TDI’s Precision Medicine Program, which helped launched Google’s Project Euphonia.
Project Euphonia uses artificial intelligence to leverage big data from the Precision Medicine Program to make ALS-affected voices understandable to Google, and eventually, to use recordings of the patients’ own voices to speak back, instead of computer-generated voices.
“No one wants to hear a computer,” said Steve Perrin, chief executive officer of ALS TDI. “Communication is such a part of our life that we take for granted. To give somebody back this ability, and their own voice, it’s transformational. Communication is king.”
Perrin said participants in the program are asked to find old recordings of their voices from before they got sick with amyotrophic lateral sclerosis, which can then be loaded into Google and used to replace the cold, emotionless sound of a computer voice.
ALS patients experience progressive muscle weakness and slowly lose the ability to perform day-to-day functions as simple as sending a text — so they often rely on voice recognition software, until even their speech becomes too difficult for the technology to handle.
“When your voice goes, it’s like a foreign language,” said Perrin.
Project Euphonia, which can be used by all kinds of patients with speech difficulties, also is programmed to understand compromised speech, which it translates into the patient’s original voice.
Seger, who is from Vancouver, Canada, said that before Project Euphonia, he typed into an app that vocalized the words, but his typing was slow and clumsy. Now, he speaks right into the Euphonia app, and it understands him.
“The Euphonia app gives me confidence that I can communicate with anyone at any time if required,” said Seger, whose successful career as a physician was taken away by a failing voice and declining word clarity.
The speech recognition tools, which were generated from thousands of ALS-affected voice recordings, will allow people to independently operate their phone or Google Home device with just their voice.
The Precision Medicine Program grew from 25 to 300 participants in 2014 largely due to funds raised through the Ice Bucket Challenge, a massive fundraiser led by the late Pete Frates that swept the globe.
Now, five years later, the program has enrolled over 600 people with ALS to gather data on voice, movement, lifestyle, medical and family histories and genetics. Google is also using the voice recordings to predict disease progression, which in some is very quick, and in others spans years.
“You can utilize a voice as a tool to put people on a disease progression rate that is way more accurate than what we currently use,” said Perrin.
Seger said, “For those with speech impairment, social isolation and loss of independence are big negatives. Communication technology like this that works as well as it does can make a huge difference in both realms.”