Because lab accuracy of clinical speech technology systems may be overoptimistic, clinical validation is vital to demonstrate system reproducibility-in this case, the ability of the PERCEPT-R Classifier to predict clinician judgment of American English /ɹ/ during ChainingAI motor-based speech sound disorder intervention.All five participants experienced statistically-significant improvement in untreated words following 10 sessions of combined human-ChainingAI treatment.These gains, despite a wide range of PERCEPThuman and human-human (F1-score) agreement, raise questions about best measuring classification performance for clinical speech that may be perceptually ambiguous.

Mispronunciation detection tools could increase treatment access for speech sound disorders impacting, e.g., /ɹ/.We show age-and-sex normalized formant estimation outperforms cepstral representation for detection of fully rhotic vs. derhotic /ɹ/ in the PERCEPT-R Corpus.Gated recurrent neural networks trained on this feature set achieve a mean test participantspecific F1-score =.81 (σx=.10,med = .83,n = 48), with post hoc modeling showing no significant effect of child age or sex.

Acoustic-to-articulatory speech inversion could enhance automated clinical mispronunciation detection to provide detailed articulatory feedback unattainable by formant-based mispronunciation detection algorithms; however, it is unclear the extent to which a speech inversion system trained on adult speech performs in the context of (1) child and (2) clinical speech.In the absence of an articulatory dataset in children with rhotic speech sound disorders, we show that classifiers trained on tract variables from acoustic-to-articulatory speech inversion meet or exceed the performance of state-of-the-art features when predicting clinician judgment of rhoticity.

We present the PERCEPT-R corpus, a labeled corpus of child speakers of American English with typical speech and residual speech sound disorders affecting rhotics.We demonstrate the utility of age-and-gender normalized formants extracted from PERCEPT-R in training support vector classifiers to predict ground-truth perceptual judgments of "rhotic" (i.e., dialecttypical) and clinical "derhotic" /ɹ/ for novel speakers (mean of participant-specific f-metrics = .83;SD = .18,N = 281).