A growing body of research suggests that social engagement can play a crucial role in buffeting age-related changes in brain. However, precise measurement and understanding of its role is a difficult challenge due to the complex and multi-faceted nature of social engagement. Thus far, social engagement and its impact on brain function has been studied largely using self-reports, whose utility are limited. The objective of this proposal is to enhance the arsenal of statistical methods available to researchers for modeling social engagement, or more broadly behavior. We propose to measure temporal changes in subjects'level of social engagement, both the size of social network and the extent of engagement, using privacy-protected analysis of their interaction, sampled from their telephone conversations. Apart from the fact that telephone conversations provides a representative sample of their interaction with the larger society for most older adults (and majority of their interaction in case of restricted mobility), current technology is primed for automatic analysis of speech over telephones, more than other kinds of free flowing conversations. This proposal brings together many needed resources -- Dr. Shafran's expertise in spoken language technology, a large clinical cohort monitored with extensive neuropsychological tests, an infrastructure for evaluating new technology, expertise in clinical psychology and epidemiology through Dr. Shafran's mentor -- in an unprecedented manner to translate the advances in language processing to clinical behavioral research. This research will form the core of a 5-year career development plan for Dr. Shafran under the mentorship of an exceptional mentor with expertise in clinical psychology and epidemiology, a plan that combines didactic and practical training in a clinical study with an ongoing research program within the unique research environment of Oregon Health &Science University. This plan will foster Dr. Shafran's development into an established independent quantitative research scientist with expertise in relevant methodology for studying the relationship between complex changes in brain function, its reflection in behavior and language use.

Public Health Relevance

The proposed K25 award is aimed at re-focusing PI's career to apply his expertise in language processing and computational modeling to study aspects of human behavior, especially those that can be inferred from spoken language. Accordingly the training plan will augment PI's expertise with a broader knowledge and deeper understanding of the relevant biology of aging and behavior so that he can effectively lead or participate in multidisciplinary research on the biology and clinical science of aging.
The specific aims of the proposed research is to develop methods to extract spoken language markers from everyday life of older adults, to understand their statistical properties and to perform association studies with respect to subjects'level of social engagement and cognitive function.

National Institute of Health (NIH)
National Institute on Aging (NIA)
Mentored Quantitative Research Career Development Award (K25)
Project #
Application #
Study Section
National Institute on Aging Initial Review Group (NIA)
Program Officer
Wagster, Molly V
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Oregon Health and Science University
Engineering (All Types)
Schools of Medicine
United States
Zip Code
Petersen, Johanna; Thielke, Stephen; Austin, Daniel et al. (2016) Phone behaviour and its relationship to loneliness in older adults. Aging Ment Health 20:1084-91
Bayestehtashk, Alireza; Asgari, Meysam; Shafran, Izhak et al. (2015) Fully Automated Assessment of the Severity of Parkinson's Disease from Speech. Comput Speech Lang 29:172-185
Stark, Anthony; Shafran, Izhak; Kaye, Jeffrey (2014) Inferring Social Nature of Conversations from Words: Experiments on a Corpus of Everyday Telephone Conversations. Comput Speech Lang 28:
Iyer, Swathi P; Shafran, Izhak; Grayson, David et al. (2013) Inferring functional connectivity in MRI using Bayesian network structure learning with a modified PC algorithm. Neuroimage 75:165-75
Stark, Anthony; Shafran, Izhak; Kaye, Jeffrey (2012) Hello, Who is Calling?: Can Words Reveal the Social Nature of Conversations? Proc Conf :112-119