Characterizing sensorimotor profiles in children with residual speech sound disorder: a pilot study.

This was a pilot study with 25 children and no control group, so changes from pre- to post-treatment cannot be separated from maturation, repeated testing, or normal development. Analyses were unbalanced: one participant was excluded entirely, and missing pre-treatment ultrasound or video left 19 pre-treatment and 25 post-treatment time points for analysis. The somatosensory task was not given at the same time for all children: 11 completed it before treatment and 14 completed it after treatment because the task and materials were still being developed and shared across sites. Tongue-shape reliability was imperfect: the MCI measure showed moderate agreement between raters, while NINFL showed only fair agreement. The ultrasound measures used only the middle tongue view and did not capture lateral bracing or tongue-palate interactions, which are relevant to /ɹ/. The somatosensory task required letter recognition and tested only the tongue tip, so it may not purely measure oral sensory skill for rhotic production. Accuracy was measured acoustically during a highly supported stimulability probe, so results may not reflect everyday speech or listener-rated generalization. The paper reports several exploratory associations in a small sample, and the Discussion notes that some patterns could be chance fluctuations.

Supplied metadata lists research support from NIH extramural and non-U.S. government sources. The article text provided does not include a conflicts-of-interest or funding declaration.

Do not read the post-treatment improvements as proof that ultrasound biofeedback works. The children also received traditional therapy, and the study had no control group, so maturation or repeated testing could explain the change. Do not use the sensory tests to guide treatment selection; they were not related to tongue-shape complexity before treatment, and the post-treatment relationships came from a small pilot.