Application of a Landmark-Based Method for Acoustic Analysis of Dysphonic Speech.
Keiko Ishikawa, Marepalli B Rao, Joel MacAuslan and 1 others
PMID 30642708WHAT IT FOUND
Landmark-based acoustic analysis distinguished dysphonic from normal speech, with a 7.24% misclassification rate.
Fewer landmarks in voiced segments and more in unvoiced segments characterized dysphonia. This method shows promise for objectively assessing speech intelligibility deficits in voice disorders.
Key findings
01Normal speakers generated more landmarks on average (64.5) than dysphonic speakers (57.94), a statistically significant difference.
02Dysphonic speech had significantly more glottal and burst landmarks and fewer syllabicity landmarks than normal speech.
03A classification tree model using syllabicity and burst landmark counts separated the groups with a 7.24% misclassification rate.
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 did not measure actual intelligibility by listeners, so the link between landmark counts and communicative ability is theoretical. Only one sentence of the Rainbow Passage was analyzed, which may not represent full speech patterns. The sample size was small and restricted to specific dysphonia etiologies. The classification model has not yet been validated on a new set of recordings.
The easy way to misread this
Do not assume that landmark counts directly measure intelligibility or that this tool is ready for clinical use. The study did not include listener ratings of intelligibility, so the relationship between these acoustic markers and how well patients are understood remains to be proven.