Þunglyndi í Tungumáli: Natural language features as a diagnostic tool for depression in older adult speakers of Icelandic

Abstract

Diagnosing depression is a difficult and time-consuming endeavor for health care professionals. Negative outcomes associated with depression, especially for older adults, highlight the importance of diagnosis. No research has yet been conducted with speakers of Icelandic using recent advances in natural language processing capable of identifying features in natural language that are correlated with depression. This exploratory case-control study investigates the potential for identifying linguistic features associated with depression in the speech patterns of older Icelandic-speaking adults. Using a small sample of two participants—one diagnosed with depression and one control—the study aims to test practical methods for data collection and analysis in Icelandic, a low-resource language. The findings are compared to studies conducted in other languages, providing preliminary insights into linguistic markers of depression, such as changes in vocal jitter, shimmer, and sentiment. While this study is not intended to establish definitive conclusions, it lays the groundwork for future research and contributes to the development of a non-invasive, efficient screening tool for depression in older Icelandic speakers.