Yorkston et al. (1990)

EBP THERAPY ANALYSIS

Comparison Learning Research

 

NOTE:  Scroll about 2/3s of the way down the page to view a description of the 4 rate control strategies and procedures.

KEY:

C = clinician(s)

P = participant(s)

pmh = Patricia Hargrove, blog developer

wpm = words per minute

 

SOURCE: Yorkston, K. M., Hammen, V. L., Beukelman, D. R., & Traynor, C. D. (1990). The effect of rate control on the intelligibility and naturalness of dysarthric speech. Journal of Speech and Hearing Disorders, 55, 550-560.

 REVIEWER(S): pmh

DATE: January 3, 2015

 

ASSIGNED GRADE FOR OVERALL QUALITY: C- (The highest possible grade for this design was B+.)

 

TAKE AWAY: This investigation is concerned with learning and should not be considered evidence of the effectiveness of an intervention. Nevertheless, learning research can provide guidance to clinicians (C). The findings from this investigation indicated that slowing the rate of speech can result in improvements in the speech of participants (Ps) with ataxic or hypokinetic dysarthria. Metered strategies are more likely to improve sentence intelligibility, although one type (Additive Rhythmic) of rhythmic rate control strategy also results in sentence intelligibility improvement. On the other hand, metered strategies tend to be associated with the more severe degrading of ratings of speech naturalness than the rhythmic approaches.

 

 

  1. What type of evidence was identified?
  • What was the type of evidence? Comparison Research–Prospective, Nonrandomized Group Design with Controls
  • What was the focus of the research? Clinically Related
  • What was the level of support associated with the type of evidence? Level = B+

                                                                                                           

  1. Group membership determination:
  • If there were groups, were participants randomly assigned to groups? No
  • If the Ps were not randomly assigned to groups, were members of groups carefully matched? Yes

                                                                    

    3.  Were experimental conditions concealed?

  • from participants? No
  • from administrators of experimental conditions? No
  • from analyzers/judges? Yes

                                                                   

  1. Were the groups adequately described? Yes

 How many participants were involved in the study?

  • total # of participant:  12
  • # of groups:  3
  • # of participants in each group: 4
  • List names of groups: Ataxic (A) Group, Hypokinetic (H) Group, Typical Speaking (TS) Group
  • Did all groups maintain membership? No. Only partial data are reported for 1 P from the A group due to a change in her medical status.

  The following variables were described or controlled:

  • age: 30-70 years
  • gender: 4f, 8m
  • first language: English
  • expressive language:
  • years post onset: 3-29 years
  • etiology:

     – A group = cerebellar degeneration, traumatic brain injury (2), tumor resection

     – H group = Parkinson’s disease (3), cerebral palsy with dystonic posturing

     – TS group = all Ps had no history of neurologic disorder:

  •  Were the groups similar before intervention began? Not Applicable
  •   Were the communication problems adequately described? Unclear
  • disorder type:

     – TS group – no reported speech disorder

– A group — pure ataxic (2), ataxic/spastic (1), ataxic/flaccid (1)

– H group – all hypokinetic

 

  1. What were the different conditions for this research?

Subject (Classification) Groups?

– A group

– H group

– TS group                                                               

Experimental Conditions?

  • rate of speech (habitual, 80% of habitual, 60% of habitual)
  • rate control strategies

– Additive Metered (AM

– Additive Rhythmic (AR)

– Cued Metered (CM)

– Cued Rhythmic (CR)

 

  1. Were the groups controlled acceptably? Yes

  

  1. Were dependent measures appropriate and meaningful? Yes                                                                                                      

The dependent measures

  • Measure #1: Speaking rate in words per minute (wpm)
  • Measure #2: Sentence intelligibility
  • Measure #3: Phoneme intelligibility
  • Measure #4: Speech naturalness

The dependent measures that are subjective are

  • Measure #2: Sentence intelligibility
  • Measure #3: Phoneme intelligibility
  • Measure #4: Speech naturalness

The dependent/ outcome measures that are objective are

  • Measure #1: Speaking rate in words per minute (wpm)

                                       

  1. Were reliability measures provided? Yes, some.

Interobserver for analyzers? Yes

  • Measure #3: Phoneme intelligibility—The investigators cited previous research reporting this information. Because they used a short version of the previously researched measure, the investigators also reported the average range of judges scores:

– Overall average range = 8.9%

– Average range for consonants = 9.6%

– Average range for vowels = 17.9%

 

  • Measure #4: Speech naturalness—The average standard deviation among the 9 judges was 0.97 points on the rating scale.

 

Intraobserver for analyzers?   Yes

  • Measure #4: Speech naturalness:

– A group = 88%

– H group = 91%

– I group = 89%

 

Treatment fidelity for investigators? Yes, kind of. However, the investigators described the accuracy of the rate control conditions. That is, they determined if Ps really spoke at 60% and 80% of their habitual rates during the slowed conditions by calculating or computing the rate of speech in each of the rate conditions and rate control strategies. Overall, the investigators determined that the computer software accurately paced the Ps rate of speech.

  

  1. Description of design:
  • The investigators compared the performance of A, H, and TS group during habitual speaking rate and during 2 slowed conditions (80% and 60% of habitual rate.)
  • The rates were slowed via computer pacing using 4 different strategies:

– Additive Metered (AM)

– Additive Rhythmic (AR)

– Cued Metered (CM)

– Cued Rhythmic (CR)

  • The dependent variables/outcome measures were sentence intelligibility, phoneme intelligibility, and speech naturalness.

 

  1. What were the results of the statistical (inferential) testing?—There was no inferential testing, only descriptive statistics.
  • Measure #1: Speaking rate in words per minute (wpm) — The investigators judged that the speaking rates were accurately paced. That is, the targets of 80% and 60% of habitual speech generally were accurately produced by the Ps.
  • Measure #2: Sentence intelligibility

     – The effect of rate control on the 2 clinical (A, H) groups: The investigators judged that as speakers reduced their speaking rate, sentence intelligibility improved using measures of mean sentence intelligibility and the charting of individual performances.

– The effectiveness of each of the 4 rate control strategies was investigated for the 2 clinical groups using the data associated with the 60% rate. The investigators determined that the 2 metered strategies (AM, CM) consistently resulted in higher scores than the rhythmic (AR, CR) strategies.

– Individual rankings of the 4 rate control strategies revealed that CM was most often the most effective strategy and CR was the least effective strategy.

– My (pmh) review of the data indicated that although one of the rhythmic strategies (AR) also resulted in marked improvements of sentence intelligibility.

  • Measure #3: Phoneme intelligibility

– The investigators reported that that phoneme intelligibility did not appear to vary (improve or decrease) as the clinical Ps’ (i.e., A and H groups) speaking rate decreased.

– Inspection of the data of individual clinical Ps revealed inconsistent responses to slowed rate: some Ps improved, some Ps regressed.

– Vowel intelligibility seemed to be particularly challenging for the clinical Ps. One common trend was observed in the A group: at slowed rates, judges tended to perceive single vowels as diphthongs.

  • Measure #4: Speech naturalness

– The investigators compared the 2 clinical groups (A and H) and the TS group.

– Overall (all Ps, rates, and rate control strategies) the lowest naturalness judgments were associated with the A group. The H group’s naturalness scores were in the middle and the best naturalness ratings were for the TS group.

– The largest decrease in naturalness ratings was for the TS group when comparing the habitual and the 60% of habitual rate.

– Although the A and H groups’ trends indicated that there were decreases in naturalness rating associated with the slowed rate, the changes were minimal.

– The investigators also explored the effectiveness of the different rate control strategies on speech naturalness. For this comparison, however, they combined the data from the metered (AM, CM) and the rhythmic (AR, CR) strategies.

– For all 3 groups of Ps, the metered strategies resulted in the poorest naturalness scores. The largest decrease in naturalness scores occurred in the TS group.

 

  1. Brief summary of clinically relevant results:
  • Slowed rate of speech resulted in improved sentence (but not phoneme) intelligibility in A and H speakers.
  • The most effective rate control strategies were metered strategies (AM, CM) although, the additive rhythmic strategy seemed pretty close to the metered strategies.
  • Metered rate control strategies were consistently poorer than rhythmic rate control strategies and the habitual rate.

 

ASSIGNED GRADE FOR QUALITY OF EXTERNAL EVIDENCE: ___C-__

 

 

 

SUMMARY OF PROCEDURES

 

PURPOSE: to investigate the effect of slowed rate and four rate control strategies on sentence intelligibility, phoneme intelligibility, and speech naturalness.

POPULATION: Ataxic dysarthria, Hypokinetic dysarthria (Parkinson’s disease, PD), and typical speakers (TS)

 

MODALITY TARGETED: expression

 

ELEMENTS/FUNCTIONS OF PROSODY TARGETED: rate

 

ELEMENTS OF PROSODY USED AS INDEPENDENT VARIABLE: rate, rhythm

 

OTHER ASPECTS OF LANGUAGE/COMMUNICATION TARGETED: sentence intelligibility, phoneme intelligibility, and speech naturalness

DOSAGE: 3 two-hour sessions (this was not an intervention; it was a learning experiment.)

 

GENERAL PROCEDURE:

 

  • There were 4 rate control strategies:

– Additive Metered (AM): The C presented the words in a targeted sentence one word at a time on a computer screen at the predetermined speaking rate. Each word was presented on the screen for the same amount of time. (C had previously shared the sentences with C so as to familiarize him/her with the sentences.)

– Additive Rhythmic (AR): The C presented the words in a targeted passage using timing one would produce in typical speech. Each word was presented on the screen for the amount of time a typical speaker would produce the word. (C had previously shared the sentences with C so as to familiarize him/her with the sentences.)

– Cued Metered (CM): C presented the entire target passage to the P on a computer screen. C cued the words at the predetermined rate by underlining each targeted word. Each word was underlined for the same amount of time

– Cued Rhythmic (CR): C presented the entire target passage on a computer screen to the P. C cued the words at the predetermined rate by underlining each targeted word. Each word was presented on the screen for the amount of time a typical speaker would produce the word.

  • The investigators identified each P’s habitual rate of speaking using a set of read stimuli. They then had Ps read other similar stimuli at slowed rates of speech (60% and 80% of the habitual rate) using the 4different rate control strategies.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: