Deep Learning-Based Analysis of Glottal Attack and Offset Times in Adductor Laryngeal Dystonia.
Ahmed M Yousef, Dimitar D Deliyski, Mohsen Zayernouri and 2 others
PMID 37977969WHAT IT FOUND
An automated tool measured vocal fold timing at the start and end of speech.
People with adductor laryngeal dystonia had longer start delays than normal speakers, but end delays did not differ. This is early technical work, not a diagnostic test.
Key findings
01Average glottal attack time was 18.95 milliseconds in the adductor laryngeal dystonia group and 14.65 milliseconds in the vocally normal group, and the difference was statistically significant with p-value less than 0.001.
02Average glottal offset time was 28.9 milliseconds in the adductor laryngeal dystonia group and 27.3 milliseconds in the vocally normal group, and the difference was not statistically significant with p-value 0.20.
03The automated method matched manual ratings closely, with correlation coefficients of 0.93 for glottal attack time and 0.91 for glottal offset time.
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.
What it does not show
The sample was very small, with four vocally normal participants and five patients with adductor laryngeal dystonia. The study did not show that glottal attack time or glottal offset time can diagnose adductor laryngeal dystonia in routine clinical practice. The automated method was compared only with manual ratings on the same recordings, not with other clinical assessment tools. The deep learning tool was trained on a small number of subjects, although frames were selected to represent varied glottal shapes and events. Recording quality varied, including brightness, contrast, blurring, and image noise, which may have affected manual and automated timing.
Declared interests
The authors declared no conflict of interest and no known competing financial interests or personal relationships that could have influenced the work.
The easy way to misread this
Do not read the longer glottal attack time as a diagnostic marker for adductor laryngeal dystonia. The study included four vocally normal participants and five patients with adductor laryngeal dystonia, and it tested an experimental automated measurement method, so it does not show clinical usefulness.
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 →