Right now, an estimated 5.2 million Americans are living with Alzheimer's disease. By 2025, that number is expected to exceed 7.1 million. By 2050, occurrence of the disease is projected to nearly triple, affecting more than 13.8 million people. Yet, there is no single diagnostic test that shows whether an individual has the disease.
"While physicians can and do make a 'clinical diagnosis' of Alzheimer's disease based on the patient's history, physical exam, and screening tests, a definitive diagnosis isn't possible without pathological analysis of the brain," explains Shawn Berkowitz, MD, director of geriatrics at UHS and associate director of Binghamton University's Southern Tier Center on Aging.
Berkowitz hopes to change that, and is working with J. David Schaffer, PhD, visiting research professor of bioengineering at Binghamton, to determine if a speech-based diagnostic test can be developed to identify Alzheimer's.
The project is a personal one for Schaffer, whose wife contracted early-onset Alzheimer's when he was a fellow at Philips Research, where he spent 25 years as a leading researcher before joining the University in 2010.
"I observed certain changes in my wife's speech patterns due to Alzheimer's, and I was aware that some of my colleagues at Philips had done work in speech analysis, particularly with speech-to-text technology, so it was sort of natural to put those two ideas together," says Schaffer.
Philips Research wasn't interested in funding a project, "so when I retired from Philips and came to Binghamton, this was one of the top items on my 'to do' list," Schaffer notes.
The desire to conduct research into detecting Alzheimer's brought Schaffer to Binghamton's Institute for Intergenerational Studies, which includes the Southern Tier Center on Aging. It was there Schaffer met Berkowitz.
"Schaffer came into the center interested in pursuing this research, and Laura Bronstein [interim dean of the College of Community and Public Affairs, professor of social work, and director of the Institute for Intergenerational Studies] and I thought it was a valuable project, as well as a great way to combine the University's bioengineering and social work departments."
According to Schaffer and Berkowitz, certain speech defects (called aphasias) are associated with different kinds of dementias, and literature suggests that some of these aphasias might have prognostic or diagnostic ability.
"From there, it's not too far a stretch to think we might be able to develop a diagnostic test for early-stage dementia based on speech samples," says Schaffer. "If that's doable," adds Berkowitz, "then it would certainly have clinical value."
Funded by a grant from the SUNY Research Foundation, the 18-month project began in January. Schaffer and Berkowitz serve as co-principals of the study, along with Bronstein and Stephen Zahorian, department chair and professor of the University's electrical and computer engineering department. In addition, the project team includes Kim Evanoski, an adjunct professor of social work, and four social work graduate students.
The study involves recording voice samples from volunteers in two distinct groups: those who have a clinical diagnosis of Alzheimer's and those with normal cognitive abilities. The groups are being matched with regard to gender, age, and years of education.
Project volunteers must first sign a consent form; a caregiver or health care proxy must provide additional consent for the volunteers who have cognitive deficiencies. The volunteers are also given a mini mental status exam, a test commonly used to measure cognitive function. They are then asked to describe an illustration that was created specifically for the study by Binghamton University art students (shown at right). Participants' responses are digitally recorded. Volunteers may speak for as little or as much time as they desire; Schaffer says the average response duration is 1 to 2 minutes. Volunteers are recorded once, and the entire process takes approximately 30 minutes. In gratitude for their participation, volunteers receive $10.
The recorded speech samples are saved as .wav files and fed into software that analyzes the speech signal, looking for pauses. The software, developed by electrical and computer engineering students under Zahorian's direction, provides information about the length of speech events and pauses, the number of pauses, pause duration, and much more. The samples are also converted into text files so the specific words the participant uses can be studied. By examining the content of the sample, the project team can calculate variables such as vocabulary richness, occurrences of repeated words, sentence complexity, idea density, and other valuable data.
Schaffer notes that each sample provides more than 100 metrics. From these data, he and Berkowitz hope to see a pattern emerge that could one day lead to a diagnostic test that could be administered in a physician's office to determine an individual's risk of developing Alzheimer's type dementia.
"I like to say that this study has two big challenges — one scientific and one engineering," says Schaffer. "The scientific challenge is: Is there a pattern and can we find it? The engineering challenge is: Can I make this automatic so a physician can run the test in the office and get a clinically useful number out the other end?"
Berkowitz and Schaffer hope to obtain 50 volunteers in each group for this study; however, recruiting volunteers with early-stage Alzheimer's disease has been difficult.
"We're looking for a specific group of volunteers," explains Berkowitz. "People who have a confirmed diagnosis of Alzheimer's-type dementia, but whose cognitive deficits aren't so severe that they can't participate effectively by giving an adequate speech sample." He adds, "We've also found that people are intimidated either by the disease process or the research; they don't want to appear that they have lost some of their cognitive faculties because Alzheimer's has such a huge stigma attached to it."
Berkowitz continues, "There is a lot riding on this project, particularly for those with Alzheimer's. There isn't a lot of movement in terms of treatments or even diagnostic potentials that have real-world applicability, so we're behind the times with this disease. And, Alzheimer's is a disease that is increasing in frequency as people live longer. It poses a significant social and financial burden to society, so we really need to make some inroads to help with diagnosis and treatment. We need to know who has the disease in order to more effectively study treatment options."
Despite the challenges of volunteer recruitment, Schaffer is optimistic about the outcome, which he hopes will catch the interest of the National Institutes of Health.
"If we get the results we're hoping for, which would be 100 percent separation between the Alzheimer's group and the control group, then it's my dream that this pilot project will lead to a longitudinal study [a longer, in-depth project with a much larger group of participants] and perhaps 10 or 20 years from now we could have a really useful tool," Schaffer says.
While this project is the first collaboration for Schaffer and Berkowitz, Binghamton University and UHS have a long history of partnership.
"From a healthcare perspective, UHS and BU are a perfect marriage," remarks Berkowitz. "The University has great resources for research and investigation, along with faculty who have dedicated their lives to looking at things from different perspectives. UHS has clinicians who work directly with patients and are committed to providing them with the very best care. Combining the two gives us the opportunity to advance knowledge and to be a center of excellence in terms of building new and innovative methods of patient care, diagnostic investigation, and developing treatment plans," he continues.
Schaffer adds that without medical research volunteers, none of this collaboration would be possible.
"I think people who volunteer to participate in medical research are the unsung heroes of medicine," he says. "We keep their names secret because ethics committees require us to do so, but I would like to see the names of all the people who have volunteered for medical research listed on a long white wall in Washington, D.C., similar to the long black wall where the names of those who died during the Vietnam War are listed. The white wall would be much, much longer, but I think as a nation we could be awfully proud of such a thing. We researchers can have all the best ideas in the world," he continues, "but without people willing to volunteer to be tested, we would get nowhere."
Volunteers with a confirmed or suspected diagnosis of Alzheimer's type dementia, who are in the early stages of the disease, are encouraged to participate in the study. To participate in the project or to get more information, contact Kim Evanoski at 607-280-1433 or firstname.lastname@example.org.
Last Updated: 9/24/13