SLPCase-ControlJournal of voice : official journal of the Voice Foundation2026

Automated Creak Detection in Spanish Speakers with and without AdLD.

M Eugenia Castro, Fermin M Zubiaur Gomar, Katherine Marks

PMID 41820120

WHAT IT FOUND

Automated creak percentage was higher in Spanish-speaking individuals with adductor-subtype laryngeal dystonia than controls and glottic insufficiency.

It was better at identifying people without dystonia than at finding people with dystonia.

Key findings

01Automated creak percentage was higher in Spanish-speaking individuals with adductor-subtype laryngeal dystonia than in controls and in individuals with glottic insufficiency, after controlling for overall dysphonia severity.

02Creak separated dystonia from controls with AUC .88 and dystonia from glottic insufficiency with AUC .73.

03The detector had high specificity (0.95) but lower sensitivity (0.45 and 0.25), so some dystonia cases may be missed.

STILL TO COME

How it was doneWhat they foundWhat it means for SLPs

Read the rest of this summary

You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.

Already have one?

What it does not show

The sample was small, with 60 participants total, and the authors call for larger work. The adductor-subtype group included 16 speakers with AdLD, 2 with AdLD and tremor, and 2 with mixed LD, so it does not test pure AdLD alone. The comparison group was glottic insufficiency, not muscle tension dysphonia, so the study does not show whether creak separates AdLD from muscle tension dysphonia in Spanish. The creak detector was trained on US-English, Finnish, Swedish, and Japanese speakers, so performance in other languages is unknown. The tool still requires MATLAB, so its clinical use is limited. Botox history varied: 7 had never received Botox and 13 had received it at least 76 days before recording. Sensitivity was lower than specificity, so some dystonia cases may be missed.

Declared interests

Katherine Marks is funded by NIH NIDCD grants R21DC023280 and R01DC009029.

The easy way to misread this

Do not use creak alone to diagnose adductor-subtype laryngeal dystonia in Spanish-speaking individuals. The paper says no single automated measure should diagnose, and the detector had lower sensitivity than specificity.

Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →