Language ENvironment Analysis (LENA) in Phelan-McDermid Syndrome: Validity and Suggestions for Use in Minimally Verbal Children with Autism Spectrum Disorder.

The sample was small and convenience-based: 18 children already in ongoing studies, so it may not represent all children with Phelan-McDermid syndrome or autism spectrum disorder. The age range was wide, and only 2 of 18 children were within the LENA validation age range of under 48 months, so accuracy findings may not apply to younger or older children in the same way. Only a small portion of the audio was transcribed by a human, and one primary transcriber coded most segments; a second rater checked only 24 five-minute segments. The selected transcription segments were chosen by vocalization count quartile, which may not reflect all daily language samples. LENA did not code whether speech was meaningful or directed at another person, so speech-like vocalizations may not equal expressive language. The study did not test whether LENA can detect meaningful change over time. Findings were based on a genetic syndrome with severe language impairment, so they may not generalise to other minimally verbal populations.

Do not treat LENA as a validated standalone language measure for minimally verbal children. It missed most human-coded child speech segments, undercounted vocalizations, and raw counts were not correlated with standard expressive language scores.