The Relationship Between Listener Comprehension and Intelligibility Scores for Speakers With Dysarthria

Abstract

Purpose-This study examined the relationship between listener comprehension and intelligibility scores for speakers with mild, moderate, severe, and profound dysarthria. Relationships were examined across all speakers and their listeners when severity effects were statistically controlled, within severity groups, and within individual speakers with dysarthria.Method-Speech samples were collected from 12 speakers with dysarthria secondary to cerebral palsy. For each speaker, 12 different listeners completed two tasks (for a total of 144 listeners), one task involved making orthographic transcriptions and one task involved answering comprehension questions. Transcriptions were scored for the number of words transcribed correctly; comprehension questions were scored on a 3-point scale according to their accuracy.Results-Across all speakers, the correlation between comprehension and intelligibility scores was non-significant when the effects of severity were factored out and residual scores were examined. Within severity groups, the relationship was significant only for the mild group. Within individual speaker groups, the relationship was non-significant for all but two speakers with dysarthria.Conclusions-Findings suggest that transcription intelligibility scores do not accurately reflect listener comprehension scores. Measures of both intelligibility and listener comprehension may provide a more complete description of the information-bearing capability of dysarthric speech than either measure alone. Keywords speech intelligibility; speech perception; dysarthria; cerebral palsy; comprehensibilityThe characterization of dysarthric speech is a topic of both clinical and theoretical importance. Although the dysarthrias are a heterogeneous group of speech disorders, one common characteristic is reduced intelligibility (Duffy, 2005; Yorkston, Beukelman, Strand, & Bell, 1999). Reduced intelligibility can have a critical impact on communication abilities, and may limit vocational, educational, and social participation. As a result, quality of life may be greatly diminished.Intelligibility refers to how well a speaker's acoustic signal can be accurately recovered by a listener. Although this definition seems simple, there are many speaker-related and listenerrelated variables that can impact how well a speech signal is deciphered. For example, research has shown that message predictability (Garcia & Cannito, 1996), message length Portions of these data were presented at the American Speech Language and Hearing Association Annual Convention, Chicago, IL (November, 2003). NIH Public AccessAuthor Manuscript J Speech Lang Hear Res. Author manuscript; available in PMC 2011 January 5. NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscript (Yorkston & Beukelman, 1981), contextual cues (Hunter, Pring & Martin, 1991), visualfacial information (Hustad & Cahill, 2003), and listener experience (Tjaden & Liss, 1995) each have the potential to affect intelligibility in significant ways. Thus, intelligibili...

Comparison of Intelligibility Measures for Adults With Parkinson's Disease, Adults With Multiple Sclerosis, and Healthy Controls

“…Intelligibility should be distinguished further from the perceptual construct of comprehensibility (Yorkston, Strand, & Kennedy, 1996). Comprehensibility, as defined by Yorkston, Beukelman, and Tice (1996), refers to how much of the acoustic speech signal a listener understands when gestures, orthographic cues, semantic cues, and other types of contextual information are available (for a discussion of differences among the perceptual constructs of intelligibility, comprehensibility, and comprehension, see a review in Hustad, 2008).…”

Combining Non-Pathological Data of Different Language Varieties to Improve DNN-HMM Performance on Pathological Speech

“…In a sense, ratings of phonemes in isolated (pseudo)words are comparable to phonemic or phonetic transcriptions, where expert transcribers are supposed to indicate, as much as possible, how speech sounds have been realized, thus approximating an articulatory description of the sounds. However, it is questionable whether discrepancies observed between such phonetic transcriptions of the realized utterances and the corresponding canonical or reference transcriptions can be taken as measures of speech intelligibility, which is supposed to indicate to what extent a given utterance has been understood by a listener [1]. A similar discussion has been going on in the field of L2 pronunciation instruction, where a distinction has been made between measures of accentedness (as opposed to nativeness) and measures of intelligibility [2,16].…”

“…Yet, both Hustad et al [1] and Beijer et al [19] argue that such word level scoring may still be quite broad, suggesting that it may be necessary to also collect judgments at even finer levels of granularity, i.e. the subword level.…”

“…According to Hustad [1] "Intelligibility refers to how well a speaker's acoustic signal can be accurately recovered by a listener". Assessments of intelligibility can be used for diagnostic purposes, but also to determine the degree of progress a patient has made.…”

Motor Deficit Severity Mediates Effects of Intellectual Function in Predicting Speech Motor Control in Children with Cerebral Palsy

“…Among children with early-onset CP (n=1268), 36% have been found to have motor speech problems and 42% have been found to have impaired communication [3]. Motor speech problems in these children can adversely affect functional communication and diminish their quality of life [4].…”