Deep learning-derived measures of sound-level accuracy in primary progressive apraxia of speech: A feasibility pipeline with descriptive evidence from two cases.

The study included only two patients with apraxia of speech, which is too small to draw any general conclusions about how the tool performs in a clinical population. The design was descriptive and exploratory, not a test of whether the tool can accurately diagnose subtypes or predict treatment response. The tool did not measure prosodic features like stress or intonation, which are critical for assessing the timing subtype of apraxia. Responses to delayed auditory feedback and rate changes were inconsistent, meaning the tool did not reliably detect changes under these common clinical conditions.

The authors declared no conflicts of interest. The study was supported by the National Institutes of Health.

Do not assume this tool is ready for clinical use or that it can reliably distinguish apraxia subtypes or predict treatment response. The findings are from only two cases, and the effects of therapy manipulations like delayed feedback were inconsistent.