Search

Purpose Auditory perceptual judgments are commonly used to diagnose dysarthria and assess treatment progress. The purpose of the study was to examine the acoustic underpinnings of perceptual speech abnormalities in individuals with Parkinson's disease (PD). Method Auditory perceptual judgments were obtained from sentences produced by 13 speakers with PD and five healthy older adults. Twenty young listeners rated overall ease of understanding, articulatory precision, voice quality, and prosodic adequacy on a visual analog scale. Acoustic measures associated with the speech subsystems of articulation, phonation, and prosody were obtained, including second formant transitions, articulation rate, cepstral and spectral measures of voice, and pitch variations. Regression analyses were performed to assess the relationships between perceptual judgments and acoustic variables. Results Perceptual impressions of Parkinsonian speech were related to combinations of several acoustic variables. Approximately 36%–49% of the variance in the perceptual ratings were explained by the acoustic measures indicating a modest acoustic perceptual relationship. Conclusions The relationships between perceptual ratings and acoustic signals in Parkinsonian speech are multifactorial and involve a variety of acoustic features simultaneously. The modest acoustic perceptual relationships, however, suggest that future work is needed to further examine the acoustic bases of perceptual judgments in dysarthria.

PurposeThis study investigated the impact of lexical characteristics on the intelligibility of speakers with Parkinson's disease (PD). Intelligibility was compared for listening in a quiet versus a noisy environment.MethodA total of 192 young listeners participated in the study, with 96 listeners listening in quiet and 96 listening in noise in which the spoken sentences were mixed with 10-talker babble. The listeners transcribed spoken sentences with target words controlled for word frequency and neighborhood density. The sentences were produced by 12 speakers with PD and 12 healthy control (HC) speakers. Transcription accuracy was compared across lexical categories between PD and HC and between the two listening conditions.ResultsTranscription accuracy was higher for high-frequency words than low-frequency words for both speaker groups in quiet and in noise. The neighborhood density effect was evident only in the noise condition with better intelligibility for words from sparse neighborhoods than from dense neighborhoods. Regardless of listening condition, the PD group was less intelligible with a significant decrease in intelligibility in noise relative to quiet.ConclusionsThe lexical properties of the words affected listeners' understanding of Parkinsonian speech. Frequent words in sparse neighborhood enhanced intelligibility, especially under adverse listening conditions. Babble noise had a detrimental impact on the intelligibility of Parkinsonian speech. The extent of the lexical effect on intelligibility increased in noise.