DementiaBank Korean Kang Corpus

Byung Ok Kang
Intelligence Information Research Division
Electronics and Telecommunications Research Institute
bokang@etri.re.kr

Byounghwa Lee
Intelligence Information Research Division
Electronics and Telecommunications Research Institute
byounghwa.lee@etri.re.kr

Participants:	37 HC, 40 MCI
Type of Study:	elicited
Location:	Korea
Media type:	audio
DOI:	doi:10.21415/PJVS-9X49

Citation information

Lee, B., Song, H. J., Park, Y.-J., & Kang, B. O. (2025). Multimodal Alzheimer’s disease recognition from image, text and audio. Scientific Reports, 15(1). https://doi.org/10.1038/s41598-025-14998-7

Project Description

Korean speech recordings and manual transcription files from 40 older adults with MCI and 37 cognitively normal older adults For each participant, the dataset consists of responses to 16 questions in the following format:

“Could you tell me what you liked most about your hometown, and any scenery or aspects you would like to be proud of or share?”
“Could you talk about the job you did for the longest time, or what you have been doing recently?”

Demographic data are available here.

Usage Restrictions

The data shall be shared solely for research purposes and shall be accessible only to researchers and institutions approved through DementiaBank’s membership review process under restricted access permissions. The data must not be used for commercial purposes. The data may be analyzed, and findings may be reported only in aggregated form. In general, the original audio recordings must not be played publicly for large audiences, including in presentations, conferences, or online postings.

Acknowledgements

This research was supported by the National Research Council of Science & Technology (NST) grant by the Korean government (MSIT) (No. CAP21000-000)