SLPOtherTrends in hearing2024

Automated Measurement of Speech Recognition, Reaction Time, and Speech Rate and Their Relation to Self-Reported Listening Effort for Normal-Hearing and Hearing-Impaired Listeners Using various Maskers.

Inga Holube, Stefan Taesler, Saskia Ibelings and 2 others

PMID 39311635

WHAT IT FOUND

Automated speech recognition accurately measured how long listeners took to respond to sentences, matching human scoring.

Reaction time tracked self-reported listening effort closely. This offers a practical way to gauge cognitive load during hearing tests without asking patients to rate their effort.

Key findings

01An automated speech recognition system produced speech-recognition scores and reaction times that were nearly identical to those obtained by a human examiner.

02The time taken to begin a verbal response decreased as the signal-to-noise ratio improved, and this measure correlated strongly with self-reported listening effort.

03Reaction time was a better predictor of listening effort than the speed of articulation during the response itself.

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 size was small (27 participants) and was a convenience sample from a previous thesis. The study used German matrix sentences, which may limit direct applicability to other languages or speech materials. Self-reported listening effort was measured on a categorical scale, which required statistical transformation (rank normalization) to compare with continuous measures like reaction time. The analysis was a re-use of existing data, not a prospectively designed study for these specific automated metrics.

Declared interests

The authors declared no conflicts of interest. Funding was provided by government initiatives in Lower Saxony, Germany, and the Volkswagen Foundation.

The easy way to misread this

Do not assume that automated speech recognition can replace clinical judgment for all populations. The study used a specific German sentence test and a small group of native speakers. The high correlation between reaction time and effort was found after statistical transformation of individual data, meaning absolute reaction times vary greatly between people.

Read it on PubMed →