In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Dave Arnold retired in his 60s and began spending time volunteering in local schools. But then he started misplacing items, forgetting appointments and losing his sense of direction. Eventually he was diagnosed with early stage Alzheimer’s.
“Hearing the diagnosis made me very emotional and tearful,” he said. “I immediately thought of all my mom had experienced.” His mother suffered with the condition for years before passing away. Over the last year, Arnold has worked for the Alzheimer’s Association as one of its early stage advisors, sharing his insights to help others in the initial stages of the disease.
Arnold was diagnosed sooner than many others. It's important to find out early, when interventions can make the most difference. One promising avenue is looking at how people talk. Research has shown that Alzheimer’s affects a part of the brain that controls speech, resulting in small changes before people show other signs of the disease.
Now, Canary Speech, a company based in Utah, is using AI to examine elements like the pitch of a person’s voice and their pauses. In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Developing the AI model
Canary Speech’s CEO, Henry O’Connell, met cofounder Jeff Adams about 40 years before they started the company. Back when they first crossed paths, they were both living in Bethesda, Maryland; O’Connell was a research fellow at the National Institutes of Health studying rare neurological diseases, while Adams was working to decode spy messages. Later on, Adams would specialize in building mathematical models to analyze speech and sound as a team leader in developing Amazon's Alexa.
It wasn't until 2015 that they decided to make use of the fit between their backgrounds. ““We established Canary Speech in 2017 to build a product that could be used in multiple languages in clinical environments,” O'Connell says.
The need is growing. About 55 million people worldwide currently live with Alzheimer’s, a number that is expected to double by 2050. Some scientists think the disease results from a buildup of plaque in the brain. It causes mild memory loss at first and, over time, this issue get worse while other symptoms, such as disorientation and hallucinations, can develop. Treatment to manage the disease is more effective in the earlier stages, but detection is difficult since mild symptoms are often attributed to the normal aging process.
O’Connell and Adams specialize in the complex ways that Alzheimer’s effects how people speak. Using AI, their mathematical model analyzes 15 million data points every minute, focusing on certain features of speech such as pitch, pauses and elongation of words. It also pays attention to how the vibrations of vocal cords change in different stages of the disease.
To create their model, the team used a type of machine learning called deep neural nets, which looks at multiple layers of data - in this case, the multiple features of a person’s speech patterns.
“Deep neural nets allow us to look at much, much larger data sets built out of millions of elements,” O’Connell explained. “Through machine learning and AI, we’ve identified features that are very sensitive to an Alzheimer’s patient versus [people without the disease] and also very sensitive to mild cognitive impairment, early stage and moderate Alzheimer's.” Based on their learnings, Canary is able to classify the disease stage very quickly, O’Connell said.
“When we’re listening to sublanguage elements, we’re really analyzing the direct result of changes in the brain in the physical body,” O’Connell said. “The brain controls your vocal cords: how fast they vibrate, the expansion of them, the contraction.” These factors, along with where people put their tongues when talking, function subconsciously and result in subtle changes in the sounds of speech.
Further testing is needed
In an initial trial, Canary analyzed speech recordings from phone calls to a large U.S. health insurer. They looked at the audio recordings of 651 policyholders who had early stage Alzheimer’s and 1018 who did not have the condition, aiming for a representative sample of age, gender and race. They used this data to create their first diagnostic model and found that it was 96 percent accurate in identifying Alzheimer’s.
Christian Herff, an assistant professor of neuroscience at Maastricht University in the Netherlands, praised this approach while adding that further testing is needed to assess its effectiveness.
“I think the general idea of identifying increased risk for cognitive impairment based on speech characteristics is very feasible, particularly when change in a user’s voice is monitored, for example, by recording speech every year,” Herff said. He noted that this can only be a first indication, not a full diagnosis. The accuracy still needs to be validated in studies that follows individuals over a period of time, he said.
Toby Walsh, a professor of artificial intelligence at the University of New South Wales, also thinks Canary’s tool has potential but highlights that Canary could diagnose some people who don’t really have the disease. “This is an interesting and promising application of AI,” he said, “but these tools need to be used carefully. Imagine the anxiety of being misdiagnosed with Alzheimer’s.”
As with many other AI tools, privacy and bias are additional issues to monitor closely, Walsh said.
Other languages
A related issue is that not everyone is fluent in English. Mahnaz Arvaneh, a senior lecturer in automatic control and systems engineering at the University of Sheffield, said this could be a blind spot.
“The system may not be very accurate for those who have English as their second language as their speaking patterns would be different, and any issue might be because of language deficiency rather than cognitive issues,” Arvaneh said.
The team is expanding to multiple languages starting with Japanese and Spanish. The elements of the model that make up the algorithm are very similar, but they need to be validated and retrained in a different language, which will require access to more data.
Recently, Canary analyzed the phone calls of 233 Japanese patients who had mild cognitive impairment and 704 healthy people. Using an English model they were able to identify the Japanese patients who had mild cognitive impairment with 78 percent accuracy. They also developed a model in Japanese that was 45 percent accurate, and they’re continuing to train it with more data.
The future
Canary is using their model to look at other diseases like Huntington’s and Parkinson’s. They’re also collaborating with pharmaceuticals to validate potential therapies for Alzheimer’s. By looking at speech patterns over time, Canary can get an indication of how well these drugs are working.
Dave Arnold and his wife dance at his nephew’s wedding in Rochester, New York, ten years ago, before his Alzheimer's diagnosis.
Dave Arnold
Ultimately, they want to integrate their tool into everyday life. “We want it to be used in a smartphone, or a teleconference call so that individuals could be examined in their home,” O’Connell said. “We could follow them over time and work with clinical teams and hospitals to improve the evaluation of patients and contribute towards an accurate diagnosis.”
Arnold, the patient with early stage Alzheimer’s, sees great promise. “The process of getting a diagnosis is already filled with so much anxiety,” he said. “Anything that can be done to make it easier and less stressful would be a good thing, as long as it’s proven accurate.”
Our daily soundscape is a cacophony of earsplitting jets, motorcycles, and construction sites. Engineers know how to eliminate and control noise, but other countries are ahead of the U.S. when it comes to keeping the quiet - with related health benefits.
Walker ultimately sold her craft equipment and returned to school to study public health. Today she is assistant professor of epidemiology and director of the Community Noise Lab at the Brown University School of Public Health. “We treat noise like a first world problem—like a sacrifice we should have to make for modern conveniences. But it’s a serious environmental stressor,” she asserts.
Our daily soundscape is a cacophony of earsplitting jets, motorcycles, crying babies, construction sites or gunshots if you’re in the military. Noise exposure is the primary cause of preventable hearing loss. Researchers have identified links between excessive noise and a heightened risk of heart disease, metabolic disorders, anxiety, depression, sleep disorders, and impaired cognition. Even wildlife suffers. Blasting oil drills and loud shipping vessels impede the breeding, feeding and migration of whales and dolphins.
At one time, the federal government had our back… and our ears. Congress passed the Noise Control Act in 1972. The Environmental Protection Agency set up the Office of Noise Abatement and Control (ONAC) to launch research, explore solutions and establish noise emission standards. But ONAC was defunded in 1981 amidst a swirl of antiregulatory sentiment.
Impossibly Loud and Unhealthy
Daniel Fink. a physician, WHO consultant, and board chair of The Quiet Coalition, a program of the nonprofit Quiet Communities, likens the effect of noise to the invisible but cumulative harm of second-hand smoke. About 1 in 4 adults in the U.S. who report excellent to good hearing already have some hearing loss. The injury can happen after one loud concert or from years with a blaring TV. Some people are more genetically susceptible to noise-related hearing loss than others.
“People say noise isn’t a big deal but it bothers your body whether you realize it or not,” says Ted Rueter, director of Noise Free America: A Coalition to Promote Quiet. Noise can chip away at your ears or cardiovascular system even while you’re sleeping. Rueter became a “quiet advocate” while a professor at UCLA two decades ago. He was plagued by headaches, fatigue and sleep deprivation caused by the hubbub of Los Angeles, he says.
The louder a sound is, and the longer you are exposed to it, the more likely it will cause nerve damage and harmful fluid buildup in your inner ear. Normal speech is 50-60 decibels (dBs). The EPA recommends that 24-hour exposure to noise should be no higher than 70 weighted decibels over 24 hours (weighted to approximate how the human ear perceives the sound) to prevent hearing loss but a 55 dB limit is recommended to protect against other harms from noise, too.
The decibel scale is logarithmic. That means 80 dB is 10 times louder than 70 dB. Trucks and motorcycles run 90 dBs. A gas-powered leaf blower, jackhammer or snow blower will cost you 100 dBs. A rock concert is in the 110 dB range. Aircraft takeoffs or sirens? 120 dBs.
Walker, the Brown professor, says that sound measurements often use misleading metrics, though, because they don’t include low frequency sound that disturb the body. The high frequency of a screeching bus will register in decibels but the sound that makes your chest reverberate is not accounted for, she explains. ‘How loud?’ is a superficial take when it comes to noise, Walker says.
After realizing the impact of noise on her own health, Erica Walker was inspired to change careers and become director of the Community Noise Lab at the Brown University School of Public Health.
Erica Walker
Fink adds that the extent to which noise impairs hearing is underestimated. People assume hearing loss is due to age but it’s not inevitable, he says. He cites studies of older people living in quiet, isolated areas who maintain excellent hearing. Just like you can prevent wrinkles by using sunscreen, you can preserve hearing by using ear plugs when attending fireworks or hockey games.
You can enable push notifications on a Smart Watch to alert you at a bar exceeding healthy sound levels. Free apps like SoundPrint, iHEARu, or NoiseTube can do decibel checks, too, but you don’t need one, says Fink. “If you can’t carry a conversation at normal volume, it’s too loud and your auditory health is at risk,” he says.
About 40 million U.S. adults, ages 20-69, have noise-induced hearing loss. Fink is among them after experiencing tinnitus (ringing or buzzing in the ears) on leaving a raucous New Year’s Eve party in 2007. The condition is permanent and he wears earplugs now for protection.
Fewer are aware of the link between noise pollution and heart disease. Piercing noise is stressful, raising blood pressure and heart rate. If you live near a freeway or constantly barking dog, the chronic sound stress can trigger systemic inflammation and the vascular changes associated with heart attacks and stroke.
Researchers at Rutgers University’s Robert Wood Johnson Medical School, working with data from the state’s Bureau of Transportation, determined that 1 in 20 heart attacks in New Jersey during 2018 were due to noise from highways, trains and air traffic. That’s 800 heart attack hospitalizations in the state that year.
Another study showed that incidence of hypertension and hardening arteries decreased during the Covid-19 air lockdown among Poles in Krakow routinely exposed to aircraft noise. The authors, comparing their pre-pandemic 2015 results to 2020 data, concluded it was no coincidence.
Mental health takes a hit, too. Chronic noise can provoke anxiety, depression and violence. Cognitively, there is ample evidence that noise disturbance lowers student achievement and worker productivity, and hearing loss among older people can speed up cognitive decline.
Noise also contributes to health disparities. People in neighborhoods with low socioeconomic status and a higher percentage of minority residents bear the brunt of noise. Affluent people have the means to live far from airports, factories, and honking traffic.
Out, Out, Damn Noise
Europe is ahead of the U.S. in tackling noise pollution. The World Health Organization developed policy guidelines used by the European Environment Agency to establish noise regulations and standards, and progress reports are issued.
Americans are relying too much on personal protective equipment (PPE) instead of eliminating or controlling noise. The Centers of Disease Control and Prevention rank PPE as the least useful response. Earplugs and muffs are effective, says Walker, but these devices are “a band-aid on a waterfall.”
Editing out noise during product design is the goal. Engineers have an arsenal of techniques and know-how for that. The problem is that these solutions aren’t being applied.
A better way to lower the volume is by maintaining or substituting equipment intended for common use. Piercing building alarms can be replaced with visual signals that flash alerts. Clanking chain and gear drives can be swapped out with belt drives. Acoustical barriers can wall off highway noise. Hospitals can soften beeping monitors and limit loudspeaker blasts. Double paned windows preserve quiet.
Editing out noise during product design is the goal. Engineers have an arsenal of techniques and know-how for that. The problem is that these solutions aren’t being applied, says Jim Thompson, an engineer and editor of the Noise Control Engineering Journal, published by the Institute of Noise Control Engineering of the USA
Engineers have materials to insulate, absorb, reflect, block, seal or diffuse noise. Building walls can be padded. Metal gears and parts can be replaced with plastic. Clattering equipment wheels can be rubberized. In recent years, building certifications such as LEED have put more emphasis on designs that minimize harmful noise.
Walker faults urban planners, too. A city’s narrow streets and taller buildings create a canyon effect which intensifies noise. City planners could use bypasses, rerouting, and other infrastructure strategies to pump down traffic volume. Sound-absorbing asphalt pavement exists, too.
Some municipalities are taking innovative measures on their own. Noise cameras have been installed in Knoxville, Miami and New York City this year and six California cities will join suit next year. If your muffler or audio system registers 86 dB or higher, you may receive a warning, fine or citation, similar to how a red-light camera works. Rueter predicts these cameras will become commonplace.
Based on understanding how metabolic processes affect noise-induced hearing loss in animal models, scientists are exploring whether pharmacological interventions might work to inhibit cellular damage or improve cellular defenses against noise.
Washington, DC, and the University of Southern California have banned gas-powered leaf blowers in lieu of quieter battery-powered models to reduce both noise and air pollution. California will be the first state to ban the sale of gas-powered lawn equipment starting 2024.
New York state legislators enacted the SLEEP (Stop Loud and Excessive Exhaust Pollution) Act in 2021. This measure increases enforcement and fines against motorists and repair shops that illegally modify mufflers and exhaust systems for effect.
“A lot more basic science and application research is needed [to control noise],” says Thompson, noting that funding for this largely dried up after the 1970s. Based on understanding how metabolic processes affect noise-induced hearing loss in animal models, scientists are exploring whether pharmacological interventions might work to inhibit cellular damage or improve cellular defenses against noise.
Studying biochemical or known genetic markers for noise risk could lead to other methods for preventing hearing loss. This would offer an opportunity to identify people with significant risk so those more susceptible to hearing loss could start taking precautions to avoid noise or protect their ears in childhood.
These efforts could become more pressing in the near future, with the anticipated onslaught of drones, rising needs for air conditioners, and urban sprawl boding poorly for the soundscape. This, as deforestation destroys natural carbon absorption reservoirs and removes sound-buffering trees.
“Local and state governments don’t have a plan to deal with [noise] now or in the future,” says Walker. “We need to think about this with intentionality.”
