MATLAB Reference Implementation and Supporting Praat Scripts (Cummins et al., 2014)

Purpose Following recent attempts to quantify articulatory impairment in speech, the present study evaluates the usefulness of a novel measure of motor stability to characterize dysarthria. Method The study included 8 speakers with ataxic dysarthria (AD), 16 speakers with hypokinetic dysarthria (HD) as a result of Parkinson's disease, and 24 unimpaired control participants. Each participant performed a series of sentence repetitions under habitual, fast, and slow speaking rate conditions. An algorithm to measure utterance-to-utterance spectro-temporal variation (UUV; Cummins, 2009) was used. Speech rate and intelligibility were also measured. Results UUV scores were significantly correlated with perceptually based intelligibility scores. There were significant differences in UUV between control speakers and the AD but not the HD groups, presumably because of differences in intelligibility in the samples used and not because of differences in pathology. Habitual speaking rate did not correlate with UUV scores. All speaker groups had greater UUV levels in the slow conditions compared with habitual and fast speaking rates. Conclusions UUV results were consistent with those of other variability indices and thus appear to capture motor control issues in a similar way. The results suggest that the UUV could be developed into an easy-to-use clinical tool that could function as a valid and reliable assessment and outcome measure.