UNIVERSITY PARK, Pa. — A four-year, $3,148,346 National Institute of Aging (NIA)-funded project aims to use computational models and psychology to study the early signs of Alzheimer’s disease and other dementias (ADRD) that may appear approximately 20 years before an official diagnosis, according to Zita Oravecz, principal investigator and associate professor of human development and family studies in the Penn State College of Health and Human Development. 

Oravecz, who is also a Penn State Institute for Computational and Data Sciences (ICDS) co-hire, will use this project to expand on her previous work related to ADRD. The research team will combine mathematical modeling and psychology to study emotions and cognition in people’s natural environments to aid in real-life, human issues such as ADRD. Her team includes co-investigators Jonathan Hakun, assistant professor of neurology, Penn State College of Medicine; Martin Sliwinski, professor of human development and family studies and director of the Penn State Center for Healthy Aging; and John Felt, assistant research professor, Center for Healthy Aging. Sharon Kim, a graduate student in the Department of Human Development and Family Studies at Penn State, will also work on this study.

“I am excited about my research,” Oravecz said. “Development of quantitative models should always be driven by real-life issues. When you have a goal this big in front of you, you’re more motivated every day to work towards it.” 

The research team will look for subtle cognitive changes in participants aged 40 to 65 through a smartphone-based study, using an app developed by the Center for Healthy Aging. Participants will complete “brain games” and surveys on their smartphones, in their natural environments, multiple times each day for two weeks at a time. The games include matching of complex symbols and remembering the location of dots on a grid. 

Oravecz’s team will develop a modern statistical toolset to extract complex signals from these repeated cognitive assessments, which will be considered in conjunction with data on the participants’ health status and aging progression.  

“This is difficult, of course, because the cognitive changes related to ADRD in midlife can be very subtle,” Oravecz said. “There’s intensive data that we have to try to break down to identify the processes indicating this subtle change. Basically, we have to extract features from the data to predict whether someone might later develop Alzheimer’s. Also, there’s a general cognitive decline that usually happens with age and we need to separate that from disease progression.”  

The researchers aim to disentangle the processes underlying cognitive change and individual differences therein using Bayesian models, which learn from data and use that prior knowledge to better predict likely outcomes. Due to the computationally demanding nature of Bayesian models, the researchers will use ICDS' supercomputer, Roar, to develop them. 

“The statistical toolkit development has several steps,” Oravecz said. “First, we formulate the mathematical/statistical models that encode our theoretical notions of what is going on behind the psychological phenomena of interest. Then, we use computational approaches to test and fit these models to data. Our approach will allow us to analyze the change process one underlying component at a time.” 

Oravecz’s team will extract important information or signals from the data and try to identify novel digital markers of subtle cognitive decline. 

Project participants will have one in-person session, which will include a blood draw to test for established biomarkers associated with Alzheimer’s risk.  

“The computation toolkit will extract digital cognitive markers from high-frequency data that we will validate against blood-based biomarkers,” Oravecz said. 

Half of the participants will complete one two-week session every six months, while the other half will do so annually, with three testing sessions for each group. This will help determine the optimal frequency for testing, Oravecz said. 

“The research could open up this kind of testing, in a way, society-wide,” Oravecz said. “Once we have our methods developed, we can monitor progress over time spans of years with these new methods ... how cognition changes, including disease progression and normative aging. We will be able to establish a risk score, and clinicians can then decide how they act on it. 

Ultimately, Oravecz said, the goal is to detect risk for Alzheimer’s reliably before symptoms are prominent enough for a definitive diagnosis.  

“This gives us a better chance, especially with current pharmaceutical and therapeutic advancements, at slowing down the disease or preventing it,” Oravecz said. “We can make a big difference.” 

Oravecz said her team also hopes to make a difference with another project aiming to better understand consensus among communities. The five-year project is funded by a $1,698,137 million grant from the John Templeton Foundation (#63364). But her work began before this project. For the past 10 years, she has used computational modeling in a cultural understanding framework to study what makes people feel loved in daily life.

“We wanted to ask people about different daily life scenarios where they might feel loved,” Oravecz said. “Is there a consensus on that? What are the signs of love in daily life? Some people may be better at picking up signs, understanding loving gestures or expressions of love while others might not pick those signs up. It’s an interesting challenge.” 

The research team, which includes co-primary investigator Saida Heshmati, assistant professor at Claremont Graduate University, and Lindy Williams, a graduate student in the Department of Human Development and Family Studies at Penn State, aims to understand how beliefs on love are shaped by personal values, religion and cultural norms, and to identify any universal experiences of love that could foster connections between cultures and people. They are also examining the link between cultural competency in love — specifically, the ability to recognize signs of love in everyday life — and psychological well-being.

Oravecz has previously published on this research, with a focus on participants within the United States. This current leg of the research is focused on participants in Japan, India, Saudi Arabia, Spain, Brazil and Kenya.  

“The main goal of this research is to increase love and happiness in the world,” Oravecz said. “Maybe if people were more aware of what makes others feel loved, there would be more love in life around the world. Research has clearly shown benefits that the more people feel loved, the happier they are. Maybe it’s something we can change and learn more about.”