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.”
(Left to right) Vickie Naylor, Bernadine Clay, and Donna Maxey read a memory prompt as they take part in the Sharing History through Active Reminiscence and Photo-Imagery (SHARP) study, September 20, 2017.
In 2020, a first-of-its kind study in JAMA linked Alzheimer’s incidence to “neighborhood disadvantage,” which is based on SDOH indicators. Through autopsies, researchers analyzed brain tissue markers related to Alzheimer’s and found an association with these indicators. In 2022, Ryan Powell, the lead author of that study, published further findings that neighborhood disadvantage was connected with having more neurofibrillary tangles and amyloid plaques, the main pathological features of Alzheimer's disease.
As of yet, little is known about the biological processes behind this, says Powell, director of data science at the Center for Health Disparities Research at the University of Wisconsin School of Medicine and Public Health. “We know the association but not the direct causal pathway.”
The corroborative findings keep coming. In a Nature study published a few months after Powell’s study, every social determinant investigated affected Alzheimer’s risk except for marital status. The links were highest for income, education, and occupational status.
Clinical trials on new Alzheimer’s medications get all the headlines but preventing dementia through policy and public health interventions should not be underestimated.
The potential for prevention is significant. One in three older adults dies with Alzheimer's or another dementia—more than breast and prostate cancers combined. Further, a 2020 report from the Lancet Commission determined that about 40 percent of dementia cases could theoretically be prevented or delayed by managing the risk factors that people can modify.
Take inactivity. Older adults who took 9,800 steps daily were half as likely to develop dementia over the next 7 years, in a 2022 JAMA study. Hearing loss, another risk factor that can be managed, accounts for about 9 percent of dementia cases.
Clinical trials on new Alzheimer’s medications get all the headlines but preventing dementia through policy and public health interventions should not be underestimated. Simply slowing the course of Alzheimer’s or delaying its onset by five years would cut the incidence in half, according to the Global Council on Brain Health.
Minorities Hit the Hardest
The World Health Organization defines SDOH as “conditions in which people are born, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life.”
Anyone who exists on processed food, smokes cigarettes, or skimps on sleep has heightened risks for dementia. But minority groups get hit harder. Older Black Americans are twice as likely to have Alzheimer’s or another form of dementia as white Americans; older Hispanics are about one and a half times more likely.
This is due in part to higher rates of diabetes, obesity, and high blood pressure within these communities. These diseases are linked to Alzheimer’s, and SDOH factors multiply the risks. Blacks and Hispanics earn less income on average than white people. This means they are more likely to live in neighborhoods with limited access to healthy food, medical care, and good schools, and suffer greater exposure to noise (which impairs hearing) and air pollution—additional risk factors for dementia.
Related Reading: The Toxic Effects of Noise and What We're Not Doing About it
Plus, when Black people are diagnosed with dementia, their cognitive impairment and neuropsychiatric symptom are more advanced than in white patients. Why? Some African-Americans delay seeing a doctor because of perceived discrimination and a sense they will not be heard, says Carl V. Hill, chief diversity, equity, and inclusion officer at the Alzheimer’s Association.
Misinformation about dementia is another issue in Black communities. The thinking is that Alzheimer’s is genetic or age-related, not realizing that diet and physical activity can improve brain health, Hill says.
African Americans are severely underrepresented in clinical trials for Alzheimer’s, too. So, researchers miss the opportunity to learn more about health disparities. “It’s a bioethical issue,” Hill says. “The people most likely to have Alzheimer’s aren’t included in the trials.”
The Cure: Systemic Change
People think of lifestyle as a choice but there are limitations, says Muniza Anum Majoka, a geriatric psychiatrist and assistant professor of psychiatry at Yale University, who published an overview of SDOH factors that impact dementia. “For a lot of people, those choices [to improve brain health] are not available,” she says. If you don’t live in a safe neighborhood, for example, walking for exercise is not an option.
Hill wants to see the focus of prevention shift from individual behavior change to ensuring everyone has access to the same resources. Advice about healthy eating only goes so far if someone lives in a food desert. Systemic change also means increasing the number of minority physicians and recruiting minorities in clinical drug trials so studies will be relevant to these communities, Hill says.
Based on SDOH impact research, raising education levels has the most potential to prevent dementia. One theory is that highly educated people have a greater brain reserve that enables them to tolerate pathological changes in the brain, thus delaying dementia, says Majoka. Being curious, learning new things and problem-solving also contribute to brain health, she adds. Plus, having more education may be associated with higher socioeconomic status, more access to accurate information and healthier lifestyle choices.
New Strategies
The chasm between what researchers know about brain health and how the knowledge is being applied is huge. “There’s an explosion of interest in this area. We’re just in the first steps,” says Powell. One day, he predicts that physicians will manage Alzheimer’s through precision medicine customized to the patient’s specific risk factors and needs.
Raina Croff, assistant professor of neurology at Oregon Health & Science University School of Medicine, created the SHARP (Sharing History through Active Reminiscence and Photo-imagery) walking program to forestall memory loss in African Americans with mild cognitive impairment or early dementia.
Participants and their caregivers walk in historically black neighborhoods three times a week over six months. A smart tablet provides information about “Memory Markers” they pass, such as the route of a civil rights march. People celebrate their community and culture while “brain health is running in the background,” Croff says.
Photos and memory prompts engage participants in the SHARP program.
OHSU/Kristyna Wentz-Graff
The project began in 2015 as a pilot study in Croff’s hometown of Portland, Ore., expanded to Seattle, and will soon start in Oakland, Calif. “Walking is good for slowing [brain] decline,” she says. A post-study assessment of 40 participants in 2017 showed that half had higher cognitive scores after the program; 78 percent had lower blood pressure; and 44 percent lost weight. Those with mild cognitive impairment showed the most gains. The walkers also reported improved mood and energy along with increased involvement in other activities.
It’s never too late to reap the benefits of working your brain and being socially engaged, Majoka says.
In Milwaukee, the Wisconsin Alzheimer’s Institute launched the The Amazing Grace Chorus® to stave off cognitive decline in seniors. People in early stages of Alzheimer’s practice and perform six concerts each year. The activity provides opportunities for social engagement, mental stimulation, and a support network. Among the benefits, 55 percent reported better communication at home and nearly half of participants said they got involved with more activities after participating in the chorus.
Private companies are offering intervention services to healthcare providers and insurers to manage SDOH, too. One such service, MyHello, makes calls to at-risk people to assess their needs—be it food, transportation or simply a friendly voice. Having a social support network is critical for seniors, says Majoka, noting there was a steep decline in cognitive function among isolated elders during Covid lockdowns.
About 1 in 9 Americans age 65 or older live with Alzheimer’s today. With a surge in people with the disease predicted, public health professionals have to think more broadly about resource targets and effective intervention points, Powell says.
Beyond breakthrough pills, that is. Like Dorothy in Kansas discovering happiness was always in her own backyard, we are beginning to learn that preventing Alzheimer’s is in our reach if only we recognized it.
Nine experts break down the biggest biotech and health breakthroughs that didn't get the attention they deserved in 2022.
New drug targets for Alzheimer’s disease
Professor Julie Williams, Director, Dementia Research Institute, Cardiff University
Genetics has changed our view of Alzheimer’s disease in the last five to six years. The beta amyloid hypothesis has dominated Alzheimer’s research for a long time, but there are multiple components to this complex disease, of which getting rid of amyloid plaques is one, but it is not the whole story. In April 2022, Nature published a paper which is the culmination of a decade’s worth of work - groups all over the world working together to identify 75 genes associated with risk of developing Alzheimer’s. This provides us with a roadmap for understanding the disease mechanisms.
For example, it is showing that there is something different about the immune systems of people who develop Alzheimer’s disease. There is something different about the way they process lipids in the brain, and very specific processes of how things travel through cells called endocytosis. When it comes to immunity, it indicates that the complement system is affecting whether synapses, which are the connections between neurons, get eliminated or not. In Alzheimer’s this process is more severe, so patients are losing more synapses, and this is correlated with cognition.
The genetics also implicates very specific tissues like microglia, which are the housekeepers in the brain. One of their functions is to clear away beta amyloid, but they also prune and nibble away at parts of the brain that are indicated to be diseased. If you have these risk genes, it seems that you are likely to prune more tissue, which may be part of the cell death and neurodegeneration that we observe in Alzheimer’s patients.
Genetics is telling us that we need to be looking at multiple causes of this complex disease, and we are doing that now. It is showing us that there are a number of different processes which combine to push patients into a disease state which results in the death of connections between nerve cells. These findings around the complement system and other immune-related mechanisms are very interesting as there are already drugs which are available for other diseases which could be repurposed in clinical trials. So it is really a turning point for us in the Alzheimer’s disease field.
Preventing Pandemics with Organ-Tissue Equivalents
Anthony Atala, Director of the Wake Forest Institute for Regenerative Medicine
COVID-19 has shown us that we need to be better prepared ahead of future pandemics and have systems in place where we can quickly catalogue a new virus and have an idea of which treatment agents would work best against it.
At Wake Forest Institute, our scientists have developed what we call organ-tissue equivalents. These are miniature tissues and organs, created using the same regenerative medicine technologies which we have been using to create tissues for patients. For example, if we are making a miniature liver, we will recreate this structure using the six different cell types you find in the liver, in the right proportions, and then the right extracellular matrix which holds the structure together. You're trying to replicate all the characteristics of the liver, but just in a miniature format.
We can now put these organ-tissue equivalents in a chip-like device, where we can expose them to different types of viral infections, and start to get a realistic idea of how the human body reacts to these viruses. We can use artificial intelligence and machine learning to map the pathways of the body’s response. This will allow us to catalogue known viruses far more effectively, and begin storing information on them.
Powering Deep Brain Stimulators with Breath
Islam Mosa, Co-Founder and CTO of VoltXon
Deep brain stimulation (DBS) devices are becoming increasingly common with 150,000 new devices being implanted every year for people with Parkinson’s disease, but also psychiatric conditions such as treatment-resistant depression and obsessive-compulsive disorders. But one of the biggest limitations is the power source – I call DBS devices energy monsters. While cardiac pacemakers use similar technology, their batteries last seven to ten years, but DBS batteries need changing every two to three years. This is because they are generating between 60-180 pulses per second.
Replacing the batteries requires surgery which costs a lot of money, and with every repeat operation comes a risk of infection, plus there is a lot of anxiety on behalf of the patient that the battery is running out.
My colleagues at the University of Connecticut and I, have developed a new way of charging these devices using the person’s own breathing movements, which would mean that the batteries never need to be changed. As the patient breathes in and out, their chest wall presses on a thin electric generator, which converts that movement into static electricity, charging a supercapacitor. This discharges the electricity required to power the DBS device and send the necessary pulses to the brain.
So far it has only been tested in a simulated pig, using a pig lung connected to a pump, but there are plans now to test it in a real animal, and then progress to clinical trials.
Smartwatches for Disease Detection
Jessilyn Dunn, Assistant Professor in Duke Biomedical Engineering
A group of researchers recently showed that digital biomarkers of infection can reveal when someone is sick, often before they feel sick. The team, which included Duke biomedical engineers, used information from smartwatches to detect Covid-19 cases five to 10 days earlier than diagnostic tests. Smartwatch data included aspects of heart rate, sleep quality and physical activity. Based on this data, we developed an algorithm to decide which people have the most need to take the diagnostic tests. With this approach, the percent of tests that come back positive are about four- to six-times higher, depending on which factors we monitor through the watches.
Our study was one of several showing the value of digital biomarkers, rather than a single blockbuster paper. With so many new ideas and technologies coming out around Covid, it’s hard to be that signal through the noise. More studies are needed, but this line of research is important because, rather than treat everyone as equally likely to have an infectious disease, we can use prior knowledge from smartwatches. With monkeypox, for example, you've got many more people who need to be tested than you have tests available. Information from the smartwatches enables you to improve how you allocate those tests.
Smartwatch data could also be applied to chronic diseases. For viruses, we’re looking for information about anomalies – a big change point in people’s health. For chronic diseases, it’s more like a slow, steady change. Our research lays the groundwork for the signals coming from smartwatches to be useful in a health setting, and now it’s up to us to detect more of these chronic cases. We want to go from the idea that we have this single change point, like a heart attack or stroke, and focus on the part before that, to see if we can detect it.
A Vaccine For RSV
Norbert Pardi, Vaccines Group Lead, Penn Institute for RNA Innovation, University of Pennsylvania
Scientists have long been trying to develop a vaccine for respiratory syncytial virus (RSV), and it looks like Pfizer are closing in on this goal, based on the latest clinical trial data in newborns which they released in November. Pfizer have developed a protein-based vaccine against the F protein of RSV, which they are giving to pregnant women. It turns out that it induces a robust immune response after the administration of a single shot and it seems to be highly protective in newborns. The efficacy was over 80% after 90 days, so it protected very well against severe disease, and even though this dropped a little after six month, it was still pretty high.
I think this has been a very important breakthrough, and very timely at the moment with both COVID-19, influenza and RSV circulating, which just shows the importance of having a vaccine which works well in both the very young and the very old.
The road to an RSV vaccine has also illustrated the importance of teamwork in 21st century vaccine development. You need people with different backgrounds to solve these challenges – microbiologists, immunologists and structural biologists working together to understand how viruses work, and how our immune system induces protective responses against certain viruses. It has been this kind of teamwork which has yielded the findings that targeting the prefusion stabilized form of the F protein in RSV induces much stronger and highly protective immune responses.
Gene therapy shows its potential
Nicole Paulk, Assistant Professor of Gene Therapy at the University of California, San Francisco
The recent US Food and Drug Administration (FDA) approval of Hemgenix, a gene therapy for hemophilia B, is big for a lot of reasons. While hemophilia is absolutely a rare disease, it is astronomically more common than the first two approvals – Luxturna for RPE65-meidated inherited retinal dystrophy and Zolgensma for spinal muscular atrophy - so many more patients will be treated with this. In terms of numbers of patients, we are now starting to creep up into things that are much more common, which is a huge step in terms of our ability to scale the production of an adeno-associated virus (AAV) vector for gene therapy.
Hemophilia is also a really special patient population because this has been the darling indication for AAV gene therapy for the last 20 to 30 years. AAV trafficks to the liver so well, it’s really easy for us to target the tissues that we want. If you look at the numbers, there have been more gene therapy scientists working on hemophilia than any other condition. There have just been thousands and thousands of us working on gene therapy indications for the last 20 or 30 years, so to see the first of these approvals make it, feels really special.
I am sure it is even more special for the patients because now they have a choice – do I want to stay on my recombinant factor drug that I need to take every day for the rest of my life, or right now I could get a one-time infusion of this virus and possibly experience curative levels of expression for the rest of my life. And this is just the first one for hemophilia, there’s going to end up being a dozen gene therapies within the next five years, targeted towards different hemophilias.
Every single approval is momentous for the entire field because it gets investors excited, it gets companies and physicians excited, and that helps speed things up. Right now, it's still a challenge to produce enough for double digit patients. But with more interest comes the experiments and trials that allow us to pick up the knowledge to scale things up, so that we can go after bigger diseases like diabetes, congestive heart failure, cancer, all of these much bigger afflictions.
Treating Thickened Hearts
John Spertus, Professor in Metabolic and Vascular Disease Research, UMKC School of Medicine
Hypertrophic cardiomyopathy (HCM) is a disease that causes your heart muscle to enlarge, and the walls of your heart chambers thicken and reduce in size. Because of this, they cannot hold as much blood and may stiffen, causing some sufferers to experience progressive shortness of breath, fatigue and ultimately heart failure.
So far we have only had very crude ways of treating it, using beta blockers, calcium channel blockers or other medications which cause the heart to beat less strongly. This works for some patients but a lot of time it does not, which means you have to consider removing part of the wall of the heart with surgery.
Earlier this year, a trial of a drug called mavacamten, became the first study to show positive results in treating HCM. What is remarkable about mavacamten is that it is directed at trying to block the overly vigorous contractile proteins in the heart, so it is a highly targeted, focused way of addressing the key problem in these patients. The study demonstrated a really large improvement in patient quality of life where they were on the drug, and when they went off the drug, the quality of life went away.
Some specialists are now hypothesizing that it may work for other cardiovascular diseases where the heart either beats too strongly or it does not relax well enough, but just having a treatment for HCM is a really big deal. For years we have not been very aggressive in identifying and treating these patients because there have not been great treatments available, so this could lead to a new era.
Regenerating Organs
David Andrijevic, Associate Research Scientist in neuroscience at Yale School of Medicine
As soon as the heartbeat stops, a whole chain of biochemical processes resulting from ischemia – the lack of blood flow, oxygen and nutrients – begins to destroy the body’s cells and organs. My colleagues and I at Yale School of Medicine have been investigating whether we can recover organs after prolonged ischemia, with the main goal of expanding the organ donor pool.
Earlier this year we published a paper in which we showed that we could use technology to restore blood circulation, other cellular functions and even heart activity in pigs, one hour after their deaths. This was done using a perfusion technology to substitute heart, lung and kidney function, and deliver an experimental cell protective fluid to these organs which aimed to stop cell death and aid in the recovery.
One of the aims of this technology is that it can be used in future to lengthen the time window for recovering organs for donation after a person has been declared dead, a logistical hurdle which would allow us to substantially increase the donor pool. We might also be able to use this cell protective fluid in studies to see if it can help people who have suffered from strokes and myocardial infarction. In future, if we managed to achieve an adequate brain recovery – and the brain, out of all the organs, is the most susceptible to ischemia – this might also change some paradigms in resuscitation medicine.
Antibody-Drug Conjugates for Cancer
Yosi Shamay, Cancer Nanomedicine and Nanoinformatics researcher at the Technion Israel Institute of Technology
For the past four or five years, antibody-drug conjugates (ADCs) - a cancer drug where you have an antibody conjugated to a toxin - have been used only in patients with specific cancers that display high expression of a target protein, for example HER2-positive breast cancer. But in 2022, there have been clinical trials where ADCs have shown remarkable results in patients with low expression of HER2, which is something we never expected to see.
In July 2022, AstraZeneca published the results of a clinical trial, which showed that an ADC called trastuzumab deruxtecan can offer a very big survival benefit to breast cancer patients with very little expression of HER2, levels so low that they would be borderline undetectable for a pathologist. They got a strong survival signal for patients with very aggressive, metastatic disease.
I think this is very interesting and important because it means that it might pave the way to include more patients in clinical trials looking at ADCs for other cancers, for example lymphoma, colon cancer, lung cancers, even if they have low expression of the protein target. It also holds implications for CAR-T cells - where you genetically engineer a T cell to attack the cancer - because the concept is very similar. If we now know that an ADC can have a survival benefit, even in patients with very low target expression, the same might be true for T cells.
Look back further: Breakthroughs of 2021
https://leaps.org/6-biotech-breakthroughs-of-2021-that-missed-the-attention-they-deserved/