SLPOtherAutism research : official journal of the International Society for Autism Research2019

How effective is LENA in detecting speech vocalizations and language produced by children and adolescents with ASD in different contexts?

Rebecca M Jones, Daniela Plesa Skwerer, Rahul Pawar and 11 others

PMID 30638310

WHAT IT FOUND

LENA did not reliably detect speech from children and adolescents with autism aged 5 to 18.

It missed many child utterances and often labelled them as adult speech, especially in older participants and when the recorder was not worn.

Key findings

01In the first clinic sample of 24 minimally verbal children and adolescents, LENA agreed with human transcription for 23.15% of target-child utterances on average.

02Agreement was much lower when the recorder was placed on the table instead of worn: 3.40%, compared with 35.01% for participants wearing it.

03In the second sample, LENA agreed for 41.66% of target-child utterances on average, missed 53.84% of them, and misclassified approximately 28% as adult speech; accuracy decreased as age increased.

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

Study 1 included only 24 participants, all minimally verbal, so it does not show how LENA performs for children with more speech. Study 2 recruited 37 families, and one participant was excluded from analysis, so the analysed sample was smaller than the recruited sample. 9 participants in study 1 could not wear the recorder, and table placement made automated detection very poor, so everyday clinic use may be unreliable. Only five girls were included in the analysis of the 15 participants who wore the device, so sex differences were not settled. The work tested automated detection of utterances, not words, grammar, or clinical outcomes, so it does not show whether LENA can guide treatment decisions.

Declared interests

The paper lists research support as N.I.H., Extramural and Non-U.S. Gov't. The supplied text does not include author conflict-of-interest declarations.

The easy way to misread this

Do not conclude that LENA can replace human transcription for speech samples from older children and adolescents with autism. It missed many child utterances and often labelled them as adult speech, especially when the recorder was not worn.

Read it on PubMed →