Comparing metrics for quantification of children's tongue shape complexity using ultrasound imaging.
Heather Kabakoff, Sam Pearl Beames, Mark Tiede and 3 others
PMID 35243947WHAT IT FOUND
Ultrasound measures of tongue shape complexity can distinguish simple vowels from complex liquids in both adults and children.
However, individual sounds often overlap, so these metrics currently identify broad patterns rather than classifying specific speech errors.
Key findings
01The Modified Curvature Index (MCI) successfully separated vowels from liquids in both adult and child data, supporting its use as a broad measure of articulatory complexity.
02Individual phonemes within complexity classes showed substantial overlap in tongue shape metrics, meaning a single sound production cannot be reliably classified by complexity alone.
03For children, incorrect /ɹ/ productions generally showed lower tongue shape complexity scores than correct productions, suggesting a link between motor differentiation and accuracy for this sound.
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 study focused on typically developing children and adults, not children with speech sound disorders, so its direct clinical applicability to disordered populations is inferred rather than tested. Data was collected from midsagittal ultrasound views only, which may miss complexity from lateral bracing or grooving seen in sounds like /ʃ/ or /t/. Individual phoneme classification was poor due to significant overlap in metric values, meaning these tools currently work for broad categories (vowels vs. liquids) but not specific sounds. The child dataset was small (17 participants), limiting the generalizability of the findings.
Declared interests
The authors declared no conflicts of interest.
The easy way to misread this
Do not use these ultrasound metrics to diagnose specific phoneme errors or classify individual speech productions. The study showed that while broad categories like vowels and liquids can be distinguished, individual sounds overlap significantly in complexity scores, making single-token classification unreliable.