Academic journal article Journal of Medical Speech - Language Pathology

Listener Effort and Response Time When Transcribing Words Spoken by Children with Dysarthria

Academic journal article Journal of Medical Speech - Language Pathology

Listener Effort and Response Time When Transcribing Words Spoken by Children with Dysarthria

Article excerpt

The relationship between an objective (listener response time) and subjective (direct magnitude estimate or DME) measure of Listener effort was determined for 28 listeners transcribing audio recordings of 40 words spoken by young children with spastic dysarthria. Words were presented in background noise at +10 dB SNR. Response time score was listeners' median response time to identify a word. DME effort score was the median DME rating assigned by listeners to a word. Relationships between accuracy scores (number of listeners who transcribed a word correctly) and each measure of listener effort wore also determined. A moderately strong correlation was found between response time and DME of listener effort scores (r = 0.733, p < .001) (words with longer response times tended to have higher effort ratings). Weaker relationships were found between accuracy and response time (r = -0.460, p = .003) and accuracy and DME effort scores (r = -0.593, p < .001) (words with higher accuracy scores tended to have shorter response times and lower effort ratings). Controlling for accuracy scores, a partial correlation, coefficient of r = 0.644 (p < .001) was found between response time and DME effort scores, indicating that response time was a significant predictor of DME, independent of word accuracy.

Keywords: listener effort; direct magnitude estimation; dysarthria; children

Longstanding interest in speech intelligibility (i.e., the degree to which a spoken message can be understood) as an index of severity and a functional outcome measure for persons with dysarthria has led to increased interest in listeners' speech perception strategies when determining what a speaker with dysarthria is saying (Klasner & Yorkston, 2005). Studies that have examined factors affecting lexical decision-making and tested hypotheses about listeners' performance based on speech perception theories (e.g., Liss, Spitzer, Caviness, & Adler, 2002) have also demonstrated that listeners' behaviours are important in understanding the impact of dysarthria on the success of spoken communication.

An aspect of listener behavior that has received increasing attention in the speech recognition literature is listener effort, which refers to the attentional requirements necessary to understand speech (Hicks & Tharpe, 2002). Most of this research has focused on processing standard speech under degraded conditions (e.g., hearing loss, noisy environments). In these studies, the mental effort that listeners expend in understanding speech has been quantified using both objective (reaction or response time) and subjective (rating scale) measures. Reaction or response time (time elapsed from presenting a stimulus to when the participant responds) in lexical decision-making tasks has been used to measure how quickly information can be processed and is considered to be an objective correlate of listener effort. Word characteristics such as frequency of occurrence and phonological neighbourhood have been shown to affect reaction time in lexical decision-making tasks. As word frequency decreases and neighborhood density increases, reaction time increases (Goldinger, Luce, & Pisoni, 1989). Reaction time is also affected by listening environment. Larsby, Hallgren, Lyxell, and Arlinger (2005) found that reaction times were longer when subjects had to make a decision based on a spoken word in background noise compared to when there was no background noise. Participants with hearing impairment had longer reaction times than those with normal hearing in both quiet and noisy conditions. Subjective measures of listener effort have been obtained using a variety of rating scales. Larsby et al. (2005) used an ordinal scale with accompanying descriptors (0 = "none at all" to 10 = "extremely great"). Fraser, Gagne, Alepins, and Dubois (2010) examined the amount of listening effort expended to understand speech in noise in several listening conditions using a 0% (no effort) to 100% (very effortful) scale. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.