Skip to content
PaperFren

Depression

Can computer programs find signs of depression in our speech?

Trifu RN, Nemeș B, Herta DC, et al. · Frontiers in psychology · 2024

Open access · cc by · source: Europe PMC

Computerized analysis of brief speech samples revealed that people with depression use more negative words, more past tense, fewer prepositions, and fewer words expressing uncertainty compared to healthy individuals.

Study at a glance

Design
Case-control — Automated linguistic analysis of 5-minute pleasurable narratives in MDD vs controls
N
N=105 · 62 MDD and 43 healthy controls
Population
Romanian adults with MDD and healthy controls
Outcome
Linguistic and content markers distinguishing MDD speech

Structured fields used in claim comparison tables when every cited study has a complete layer.

Key findings

Depressed participants spoke using more past tense verbs, negative emotion words, anxiety words, auxiliary verbs, conjunctions, and absolute negations. Conversely, they used significantly fewer positive emotion words, prepositions, and tentative terms like 'maybe' or 'perhaps' compared to the control group. They also talked less about leisure activities and more about home and money, while using shorter sentences that required more periods.

Methodology

The researchers compared 62 Romanian patients diagnosed with major depressive disorder against 43 healthy controls. Each participant recorded a 5-minute spoken narrative about a hobby or something pleasurable, which was transcribed and analyzed using automated Romanian text-analysis software. They measured depression severity using clinical rating scales and evaluated differences in grammar, emotion words, and content themes.

Limitations

Because this was a cross-sectional study, it cannot prove that these specific language patterns cause depression or that depression directly causes them. The study relied on a relatively small sample of 62 patients and 43 controls, and did not track the effects of antidepressant medications, which can significantly alter energy levels and speech patterns. Additionally, since Romanian is a language where pronoun subjects are often dropped, these specific findings regarding pronoun usage might not apply to English speakers.

How this study connects

Role on claims

Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.

  • Speech patterns can carry linguistic markers that differ in depression.

    Spoken language patterns of individuals with depression feature increased past-tense verbs, negative emotion words, auxiliary verbs, and absolute negations, alongside fewer positive emotion and tentative terms.

    Evidence for the claim as stated.

  • Support can associate with better mood via lower perceived stress—related but not MDD diagnosis.

    Family/partner support associated with better mood and less anxiety/depression with mediation through lower perceived stress in a community COVID sample—useful context for distress pathways, not a clinical MDD assay.

    Scope note — different construct — community distress/support vs clinical speech markers

    Limits the claim's scope: a different population, assay, or outcome.

  • Linguistic markers, emotion differentiation, working-memory affect, and social cognition are all MDD-adjacent assays that should not be averaged into one “depression cognition score.”

    Evidence for the claim as stated.

Open questions

Tensions this paper is part of

From concept pages' “where studies disagree.” Disagreement means the same question; scope means different assays, populations, or outcomes.

History

When this study was placed

Dated entries from the concept change log — when this paper was added or removed as support, challenge, or qualifier on a claim.

  1. 2026-09-14

    Placed as a scope qualifier on Major Depressive Disorder

    Family/partner support associated with better mood and less anxiety/depression with mediation through lower perceived stress in a community COVID sample—useful context for distress pathways, not a clinical MDD assay.

  2. 2026-09-14

    Placed as supporting evidence on Major Depressive Disorder

    Linguistic markers, emotion differentiation, working-memory affect, and social cognition are all MDD-adjacent assays that should not be averaged into one “depression cognition score.”

Related papers in this topic

Same topic cluster — not a recommendation engine.