Dataset Overview

The Veteran Healthy Aging Dataset (VetAge) builds on publicly available oral history recordings from the Library of Congress Veterans History Project, curated with additional processing including speaker diarization, feature extraction, and links to mortality data.

The dataset was developed to explore whether vocal cues can predict age and remaining life expectancy in older adults. Analyses using VetAge showed that older-sounding voices were linked to shorter life expectancy when chronological age was held constant, showing the potential of speech as a noninvasive indicator of aging. 

Researchers using this dataset are kindly asked to cite the following publication, which first introduced and analyzed the data:

Yunting Yin, Douglas William Hanes, Steven Skiena, Sean A P Clouston, Quantifying Healthy Aging in Older Veterans Using Computational Audio Analysis, The Journals of Gerontology: Series A, Volume 79, Issue 1, January 2024, glad154, https://doi.org/10.1093/gerona/glad154

Bibtex:

@article{VetAge,

    author = {Yin, Yunting and Hanes, Douglas William and Skiena, Steven and Clouston, Sean A P},

    title = {Quantifying Healthy Aging in Older Veterans Using Computational Audio Analysis},

    journal = {The Journals of Gerontology: Series A},

    volume = {79},

    number = {1},

    pages = {glad154},

    year = {2023},

    month = {06},

    issn = {1758-535X},

    doi = {10.1093/gerona/glad154},

    url = {https://doi.org/10.1093/gerona/glad154},

    eprint = {https://academic.oup.com/biomedgerontology/article-pdf/79/1/glad154/54705524/glad154.pdf},

}