Funded by a seed grant from the Penn State College of Information Sciences and Technology (IST), two College of Nursing faculty members are working on an interdisciplinary project that they hope will help improve early identification of Alzheimer’s disease.
Assistant Professors Nikki Hill and Jacquie Mogle are collaborating with Professor Prasenjit Mitra in IST to build efficient algorithms for analyzing qualitative data from in-person interviews using deep learning techniques. These algorithms will help to identify aspects of subtle memory problems among older adults that indicate eventual transition to Alzheimer’s disease.
“Those affected by Alzheimer’s disease typically perceive gradual changes in memory and thinking before clinicians can detect a decline through testing,” Mogle explained. “Gathering in-depth characterizations of personal experiences from interviews is difficult because the data is expensive to collect and analyze. Deep learning techniques via text mining could possibly reduce the resource burden.”
The study will analyze transcripts from interviews with older adults using both human and computer-assisted coding. Results will be examined for consistency across methods.
“Computer-assisted coding of extensive textual data may decrease the time and cost associated with human coding,” said Mogle. “It may also provide new ways to examine individual change over time, which would aid in the early detection of Alzheimer’s disease.”