When Shena Downs left her nearly 1-year-old daughter with her mother several years ago, she returned home to find her baby without a diaper. It was one of the first signs of her mother’s forgetfulness. 

Downs’ mother was later diagnosed with dementia from Alzheimer’s disease (AD), which has progressively worsened over the past seven years. 

“It’s hard,” Downs said. “Now she doesn’t remember her kids; she walks right by us.” 

Downs, who co-owns Hummingbird Island Cuisine with her husband in downtown Wilmington, didn’t hesitate when Roxanne Williams, a community outreach coordinator for the University of Delaware’s Resilient Cognitive Aging Lab (RECALL), approached her about her restaurant serving as a critical community outreach point for a study on dementia from AD. 

“Earlier detection could slow disease progression and lead to a longer quality of life,” Downs said. “I want to be around for my daughters.”

A new approach to earlier detection

Memory loss, as Downs’ mother experienced, is often considered one of the earliest signs of AD. However, Alyssa Lanzi, assistant professor of communication sciences and disorders (CSCD) in the College of Health Sciences (CHS) and co-director of the RECALL Lab, believes language could be an even earlier marker.

“It is not easy to capture mild declines in memory, and research shows that language, including word-finding challenges and the types of words used, may be more sensitive to earlier changes,” Lanzi said.  

Lanzi and Brian MacWhinney, a professor of cognitive psychology at Carnegie Mellon University, have been awarded a $3.7 million RF1 grant from the National Institute on Aging (NIA) within the National Institutes of Health (NIH) to expand and refine DementiaBank, an open-access database of language samples for the study of language as an early behavioral marker of AD. If successful, this research could pave the way for earlier interventions. 

“Identifying these individuals as early as possible gets them into preventive treatments sooner,” Lanzi said.

The study builds on pilot data gathered by Anna Saylor, a third-year doctoral student in the communication sciences and disorders doctoral program.

“We know a lot about how language develops in childhood but not much about how it changes in older adults,” Saylor said. “Our data suggest subtle language changes might signal future cognitive decline.”

Faith Stagge, also a third-year doctoral student in the CSCD doctoral program, is conducting a discourse analysis for people at risk for dementia.  

“We must learn more about how older adults communicate because we don’t necessarily know what’s normal versus not normal,” Stagge said. “Early research shows those at risk may repeat more words or pause more, so we need to obtain more data to determine how and if that relates to their cognition.” 

A need for better data

To explore these changes on a larger scale, Lanzi is collaborating with MacWhinney, who founded TalkBank, an open science database for spoken language research. Within TalkBank is DementiaBank, a shared database of multimedia interactions for studying communication in dementia. However, DementiaBank is outdated and limited in demographics, and the quality and rigor of the data need improvement. 

Lanzi is seeking to change that. Her five-year study seeks 300 older adults aged 60-90 nationwide from underrepresented backgrounds or populations vulnerable to health disparities. 

“Current DementiaBank data is representative of Caucasians of a higher socioeconomic status,” Lanzi said. “We must intentionally recruit people who are at the greatest risk — for example, adults who are Black, Asian, Hispanic, Latin and those living in rural areas.”