Tosto et al. (2011)

June 18, 2015

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

 

Key:

 

AD = Altzheimer’s disease

NA = not applicable

NT = neurotypical

P = participant

pmh = Patricia Hargrove, blog developer

WNL = within normal limits

 

 

SOURCE: Tosto, G., Gasparini, M., G.L. Lenzi, G. L., & Bruno, G. (2011). Prosodic impairment in Altzheimer’s disease: Assessment and clinical relevance. Journal of Neuropsychiatry and Clinical Neuroscience, 23 (2), E21-E23.

Journal Address: http://neuro.psychiatryonline.org

REVIEWER(S): pmh

 

DATE: June 13, 2015

ASSIGNED GRADE FOR OVERALL QUALITY: D (Based on the case study design, the highest possible grade was D+.)

 

POPULATION: Altzheimer’s disease (AD); Adult

 

PURPOSE: To describe comprehension and production of emotional prosody in an Italian male with mild, early onset AD.

 

INSIGHTS ABOUT PROSODY:

  • This case study diverges from the clinical literature on prosody in dementia in which the expectation is that

– problems with the comprehension of emotional prosody appear in the early stages of AD and then remain stable throughout the mild and moderate stages and

– problems with the production and imitation of emotional prosody are minimal during the early stages of AD but during the moderate stage of AD, significant problems with the production and imitation of emotional prosody begin to emerge.

  • In this case study of an adult Italian speaking male diagnosed being in the early stages of early onset AD, the participant (P) displayed significant impairment in the production and comprehension of emotional prosody.
  • The investigators recommend that clinicians (C) should carefully monitor the emotional prosody skills of Ps with AD and consider emotional prosody skills when working with AD individuals and their families.

 

 

  1. What type of evidence was identified? Case Study
  1. Group membership determination:
  • If there were groups of participants were members of groups matched? Not Applicable (NA), there was only one P.
  1. Was participant’s communication status concealed?
  • from participant? No
  • from assessment administrators? No
  • from data analyzers? No

                                                                    

 

  1. Was the participant adequately described? Yes. I would have liked more information but to be fair to the investigators I rated this question as “yes” because this was a letter to the Editor and was only 3 pages long.

 

How many participants were involved in the study? 1

The following variables were described:

  • age: 55 years
  • gender: male
  • cognitive skills: 29/30 on the Mini-Mental State Exam; within normal limits (WNL)
  • neurological exam: WNL
  • diagnosis: early onset AD
  • MRI results: cortical/subcortical atrophy of fronto-parietal areas
  • PET scan results: hypometabolism in the right frontal-temporal-parietal areas, left posterior parietal areas, and basal ganglia
  • results of neuropsychological testing:

– widespread, mild cognitive impairment,

     – more impaired on tasks sensitive for both right-hemisphere and executive function

  • educational level of P: may be 14 years

 

– Were the communication problems adequately described? No

  • disorder type: AD

 

 

  1. What were the different conditions for this research?

                                                                                                             

– Subject (Classification) Groups? No. There was only 1 P. ____

If yes, list:

Experimental Conditions? No

 

– Criterion/Descriptive Conditions? Yes

  • Experimental Protocol –Melodic Alternation [previously administered to a group of neurotypical (NT) adults: 10 Ps; same gender; mean age = 54.5; mean education: 14 years]

– Task 1: Auditory Verbal Task: P listened and identified the same neutral sentence produced with sad, happy, angry, surprised, and neutral prosody

– Task 2: Visual Task: P viewed facial expressions depicting sad, happy, angry, surprised, and neutral emotions and identified each

– Task 3: Repetition Task: using the sentences from Task 1, P imitated the same sentence using the emotional prosody modeled in the sentence. Independent judges listened to the models and the imitation to judge accuracy.

– Task 4: Production Task: The investigators presented P with a sentence and directed him/her to produce the sentence with sad, happy, angry, surprised, and neutral intonations.

 

  1. Were the groups controlled acceptably? NA

 

 

  1. Were dependent measures appropriate and meaningful? Yes

 

–   The dependent measures were

  • Dependent Measure #1: Performance on Task 1: Auditory Verbal Task
  • Dependent Measure #2: Performance on Task 1: Visual Task
  • Dependent Measure #3: Performance on Task 1: Repetition Task
  • Dependent Measure #4: Performance on Task 1: Production Task

 

All the dependent measures were subjective.

 

None of the dependent/ outcome measures were objective.

                                         

 

  1. Were reliability measures provided?

                                                                                                            

– Interobserver for analyzers? Yes

  • Dependent Measure #3: Performance on Task 1: Repetition Task—K index = 0.86
  • Dependent Measure #4: Performance on Task 1: Production Task—-K index = 0.93

 

Intraobserver for analyzers? No

 

– Treatment/Procedural fidelity for investigators? No

 

  1. Description of design:
  • In this case study, a 55-year-old Italian male, who had been diagnosed as being in the early stages of early-onset AD, was administered 4 tasks to assess his emotional prosody skills.
  • The 4 tasks were

– Task 1: Auditory Verbal Task: P listened to and identified the same neutral sentence produced with sad, happy, angry, surprised, and neutral prosody

– Task 2: Visual Task: P viewed facial expressions depicting sad, happy, angry, surprised, and neutral emotions and identified each

– Task 3: Repetition Task: using the sentences from Task 1, P imitated the same sentence using the emotional prosody modeled in the sentence. Independent judges listened to the models and the imitation to judge accuracy.

– Task 4: Production Task: The investigators presented P with a sentence and directed him to produce the sentence with sad, happy, angry, surprised, and neutral intonations. Independent judges listened to the production to judge accuracy.

  • The 4 tasks, Melodic Alternation, had previously been administered to a group of neurotypical (NT) adults: 10 Ps; same gender; mean age = 54.5; mean education: 14 years. This provided comparison data.

 

 

  1. What were the results of the inferential statistical testing? NA

 

  1. What were the results of the correlational statistical testing? Correlational statistics were presented only of the reliability data for measures #3 and #4. (See item #8.)

 

  1. What were the results of the descriptive analysis?
  • Dependent Measure #1: Performance on Task 1: Auditory Verbal Task—P correctly identified 31 of 60 trials (NT control group identified 58/60.) The emotional states that exhibited most errors were neutral, happy, and sad. The P also frequently confused these 3 states.

 

  • Dependent Measure #2: Performance on Task 1: Visual Task—P correctly identified 73% of the trials (NT control group identified correctly 92%. The emotional states that exhibited most problems were neutral and sad.

 

  • Dependent Measure #3: Performance on Task 1: Repetition Task—Overall, P incorrectly imitated prosody 33% of the time. The poorest imitation rate was for happiness (83% errors); the best imitation rates were for surprise (no errors) and neutral.

 

  • Dependent Measure #4: Performance on Task 1: Production Task—All the production trails were associated with some failure. Happy (100% error rate) and angry (83% error rate) were the most challenging emotions to produce acceptably.

 

Advertisements

Samuelsson et al. (2011)

June 10, 2015

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

 

Key:

 

NA = not applicable

P = Participant

pmh = Patricia Hargrove, blog developer

SLI = Specific Language Impairment

TD = Typically Developing

 

 

SOURCE: Samuelsson, C., Reuterskiöld, C., Nettelblatt, U., & Sahlén, B. (2011.) Production and perception of metrical patterns in Swedish children with language impairment. Logopedics Phonatrics Vocology, 36, 1-11.

 

REVIEWER(S): pmh

 

DATE: June 6, 2015

ASSIGNED GRADE FOR OVERALL QUALITY:  B (The highest possible grade based on the design of the investigation was B+.)

 

POPULATION: Specific Language Impairment (SLI), Swedish; Children

 

PURPOSE: To investigate the perception and production of metrical patterns in Swedish children diagnosed with language impairment.

 

INSIGHTS ABOUT PROSODY:

  • Swedish children with SLI more frequently omitted unstressed syllable than TD peers.
  • There may be 3 subgroups of prosody among children with SLI:

– stronger perception/comprehension than production/imitation,

– stronger production/imitation than perception/comprehension, and

– equivalent production/imitation and perception/comprehension.

 

 

  1. What type of evidence was identified? Prospective, Nonrandomized Group Comparison Design
  1. Group membership determination:
  • If there were groups of participants, were members of groups matched? Yes
  • The matching strategy involved selecting participants in the two groups were close in age (preschoolers.)
  1. Was participants’ communication status concealed?
  • from participants? No
  • from assessment administrators? Unclear
  • from data analyzers? Unclear

                                                                    

 

  1. Were the groups/participants adequately described? Yes

How many participants (Ps) were involved in the study?

  • total # of participants: 52
  • was group membership maintained throughout the experiment? Yes
  • # of groups: 2
  • List names of groups: SLI and typically developing (TD)
  • # of participants in each group: SLI = 27 (but see genders- p.4, reported 28 Ps); TD = 25

           

– The following variables controlled:              

  • cognitive skills: at least 78 nonverbal IQ (SLI); teacher reported normal development (TD)
  • language skills: SLI Ps were diagnosed with language impairment by a Swedish speech-language pathologist (SLP)
  • educational level of clients: Preschool
  • hearing: both groups passed a pure tone screening (25dB at 500, 1000, 2000, and 4000 Hz)

– The following variables were described:

  • age: mean age = 4:10 (TD); mean age = 5:4 (SLI)
  • gender: 10f; 18m (SLI); TD not reported

 

– Were the communication problems adequately described? Yes

  • disorder type:

     – 50% of the SLI group was diagnosed with severe language impairment in all domains of language

– 25% of the SLI group was diagnosed with moderate, expressive language impairment, particularly with respect to grammar and phonology

– 25% of the SLI group was diagnosed with mild phonological impairment

 

 

  1. What were the different conditions for this research?
  • Subject (Classification) Groups? Yes, there were 2 groups: SLI and TD
  • Experimental Conditions? No
  • Criterion/Descriptive Conditions? Yes, there were 3 tasks, all of them were in Swedish:
  • Imitation of phrases that differed on the placement of stress on the verb particle or on the prepositional phrase (iambic vs trochaic stress.) The examiner provided pictures to support the different sentences.
  • Comprehension task in which the participant (P) hears a direction and points to the appropriate picture of a set of 3 (2 contrasting sentences and a foil). The directions were identical sentences that differed only in the placement of stress on the verb particle or on the prepositional phrase (iambic vs trochaic stress.)
  • Imitation of phrases with an indefinite article (before the stressed syllable) or a definite form of a noun (after the stressed syllable. This contrasted differed iambic vs trochaic stress. The examiner provided pictures to support the different sentences.

 

  1. Were the groups controlled acceptably? Yes

 

 

  1. Were dependent measures appropriate and meaningful? Yes
  • The dependent measures were

Dependent Measure #1: Overall number of unstressed syllables omitted across all tasks

Dependent Measure #2: Proportion of prestressed syllables omitted in content words versus the proportion of omitted copulas

Dependent Measure #3: Comparing the same morpheme in different positions (stressed versus unstressed) and with different functions (verb particle versus preposition)

Dependent Measure #4: Relative strength of performance on imitated versus comprehension tasks concerned with phrasal stress.

Dependent Measure #5: Performance of Groups from measure 4 on

  • imitation of total unstressed syllables
  • perception of total unstressed syllables
  • imitation of prestressed syllables

Dependent Measure #6: Relationship between linguistic variables and nonverbal IQ among SLI Ps

  • All the dependent measures were subjective:.

 

  • None of the dependent/ outcome measures were objective?

                                         

 

  1. Were reliability measures provided?

 

  • Interobserver for analyzers? Yes. Overall inter-rater reliability for imitated/production tasks was 95%. Differences between raters were resolved by consensus.

 

  •  Intraobserver for analyzers? No

 

  • Treatment/Procedural fidelity for investigators? No

 

 

  1. Description of design:
  • The investigators administered 3 tasks to two groups of Swedish speaking children (SLI and TD groups.)
  • Prior to the administration of the tasks, the SLI Ps had been assessed using the Swedish Test of Language Comprehension, T.R.O.G. (a comprehension test), Lund Test of Phonology and Grammar (expressive), ORIS (oral motor skills), and a test of nonverbal IQ.
  • The 3 tasks were

– Imitation of phrases that differed on the placement of stress on the verb particle or on the prepositional phrase (iambic versus trochaic stress.) The examiner provided pictures to support the different sentences.

– Comprehension task in which the participant (P) hears a direction and points to the appropriate picture of a set of 3 (2 contrasting sentences and a foil). The directions were identical sentences that differed only in the placement of stress on the verb particle or on the prepositional phrase (iambic versus trochaic stress.)

– Imitation of phrases with an indefinite article (before the stressed syllable) or a definite form of a noun (after the stressed syllable. This contrasted differed iambic versus trochaic stress. The examiner provided pictures to support the different sentences.

  • The investigators used inferential (nonparametric) and correlational statistics to analyze the data.

 

  1. What were the results of the inferential statistical testing
  • Significant comparisons are p ≤ 0.05.

Dependent Measure #1: Overall number of unstressed syllables omitted across all tasks- Ps with SLI omitted significantly more unstressed syllables and TD Ps.

Dependent Measure #2: Proportion of prestressed syllables omitted in content words versus the proportion of omitted copulas by SLI group—no significant differences

Dependent Measure #3: Comparing the same morpheme in different positions (stressed versus unstressed) and with different functions (verb particle versus preposition) by the SLI group—no significant difference

Dependent Measure #4: Relative strength of performance on imitated versus comprehension tasks concerned with phrasal stress— not analyzed statistically. See descriptive analysis Item #12 in which 3 groups (A, B, C) of Ps with SLI were identified.

Dependent Measure #5: Performance of Groups from measure 4 (see also item #12) on

  • imitation of total unstressed syllables: Group A significantly poorer than other groups.
  • perception of total unstressed syllables: Group B significantly poorer than other groups.
  • imitation of prestressed syllables: Group A significantly poorer than other groups.

Dependent Measure #6: Relationship between linguistic variables and nonverbal IQ among SLI Ps: This measure was analyzed inferentially and correlationally. See item #11 for the correlational analysis.

     – There were no significant differences among the 3 SLI groups (i.e., Groups A, B, and C) with respect to linguistic and nonlinguistic measure.

– What were the statistical test used to determine significance?

  • Mann-Whitney U
  • Kruskal-Wallis
  • Post-hoc analysis by Siegel and Castellan

 

Were evidence-based measures provided: No

– Were confidence intervals provided? No

 

 

  1. What were the results of the correlational statistical testing?

 

– Correlations that od p ≤ 0.05) are considered significant:

 

  • Dependent Measure #6: Relationship between linguistic variables and nonverbal IQ among SLI Ps: This measure was analyzed inferentially and correlationally. See item #10 for the inferential analysis.

     – Overall SLI (i.e., combined Groups A, B, and C) as well as Groups A and B individually—no significant correlation between perception and imitation of stress patterns

     – For Group C there was a significant correlation between perception and imitation of stress patterns (r = 0.85.)

correlationally. See item #10 for the inferential analysis.

     – Overall SLI (i.e., combined Groups A, B, and C), there were no significant correlations for the number of unstressed free morphemes and oral motor skills.

     – Overall SLI (i.e., combined Groups A, B, and C), there was a significant correlation for the number of unstressed free morphemes and receptive language (r = 0. 37.)

 

  1. What were the results of the descriptive analysis?

Dependent Measure #4: Relative strength of performance on imitated versus comprehension tasks concerned with phrasal stress:

  • The investigators identified 3 subgroups within the SLI group.
  • NOTE : throughout the manuscript the investigator used the term “production” instead of “imitation.” I will continue with the term “imitation” because it is more specific. Also, some of the information provided here was found in the results or discussion section.

– GROUP A: Ps who were stronger on the perception/comprehension tasks than on the imitative task. (Criterion: 2 points better on perception/ comprehension tasks than on the imitation task.) These Ps had problems with all aspects of expressive language performance. This also was the smallest group.

– GROUP B: Ps who were stronger on the imitative task than on the perception/comprehension tasks (Criterion: 2 points better on the imitation task than on the perception/ comprehension tasks.) These Ps had more trouble with overall language comprehension than the other 2 groups.

– GROUP C: Ps whose performance on the perception/comprehension tasks and the imitative task were similar. (Criterion: Scores on the perception/ comprehension tasks and the imitative task were the same or within one point of one another.) Although all these Ps had nonverbal IQs of at least 78, as a group their nonverbal IQs were lower than the o


Heikkinen et al. (2010)

April 23, 2015

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

 

Key:

 

AS = Asperger’s syndrome

NA = not applicable

P = participant or patient

pmh = Patricia Hargrove, blog developer

TYP = typically developing

WNL = within normal limits

 

 

SOURCE: Heikkinen, J., Jansson-Verkasalo, E., Toivanen, J., Kalervo, S., Väyrynen, E., Moilanen, I., & Seppänen, T. (2010). Perceptions of basic emotions from speech prosody in adolescents with Asperger’s syndrome. Logopedics Phonatrics Vovology, 35, 113-120

REVIEWER(S): pmh

 

DATE: April 20, 2015

ASSIGNED GRADE FOR OVERALL QUALITY: B

 

POPULATION: Asperger’s syndrome (AS); adolescents

 

PURPOSE: To investigate the ability of adolescents diagnosed with AS to perceive basic emotions from Finnish speech prosody

 

INSIGHTS ABOUT PROSODY:

  • Finnish adolescents diagnosed with AS performed similarly to TYP peers on a task requiring them to discriminate which prosody signals basic emotion in a read passage.

 

  1. What type of evidence was identified? Prospective, Nonrandomized Group Comparison Design
  1. Group membership determination:
  • If there were groups were members of groups matched? Yes
  • The matching strategy involved matching on chronological age.
  1. Was participants’ communication status concealed?
  • from participants? No

                                                                    

  • from assessment administrators? No
  • from data analyzers? No

                                                                    

 

  1. Were the groups/participants adequately described? Yes

– How many participants were involved in the study?

  • total # of participants: 27
  • was group membership maintained throughout the experiment? Yes
  • # of groups: 2
  • List names of groups: AS; typically developing peers (TYP)
  • # of participants in each group: AS = 12; TYP = 15

           

The following variables were described (except where listed as controlled):           

  • age: mean age   14.5 (AS); 14.3 (TYP)
  • gender: 9m, 3f (AS); 8m, 7f [controlled variable]
  • cognitive skills: only AS scores were reported:

     – total IQ mean = 106

     – total IQ range = 84 to 119

     – verbal IQ mean = 107

     – performance IQ mean = 105

  • attention deficits: exclusion criteria
  • language deficits: exclusion criteria
  • learning problems: none reported (TYP)
  • psychiatric diagnosis: exclusion criteria
  • hearing: within normal limits (WNL)
  • mental illness: none reported (TYP)
  • first language: Finnish
  • educational level of clients: primary school (TYP);

 

– Were the communication problems adequately described?

  • disorder type: all AS Ps were diagnosed with AS.

 

 

  1. What were the different conditions for this research?
  • Subject (Classification) Groups? Yes:

Asperger’s syndrome group (AS)

— Typically developing group (TYP)

                                                               

  • Experimental Conditions? Yes:

— Actors read sentences representing different emotional states:

  • Neutral (54 sentences)
  • Sad (18 sentences)
  • Happy (18 sentences)
  • Angry (18 sentences)

 

– Criterion/Descriptive Conditions? No

 

  1. Were the groups controlled acceptably? Yes

 

 

  1. Were dependent measures appropriate and meaningful? Yes

– The dependent measures were

  • Dependent Measure #1: Number of correct perceptions of neutral sentences
  • Dependent Measure #2: Number of correct perceptions of sad sentences
  • Dependent Measure #3: Number of correct perceptions of angry sentences
  • Dependent Measure #4: Number of correct perceptions of happy sentences

 

All the dependent measures were subjective.

 

– None of the dependent/ outcome measures were objective.

                                         

 

  1. Were reliability measures provided?
  • Interobserver for analyzers? No
  • Intraobserver for analyzers? No
  • Procedural fidelity for investigators? No

 

  1. Description of design:
  • Professional actors read the same passage several times in Finnish. Different emotions were depicted in the readings: neutral, sad, angry, and happy.
  • Ps listened to the passage using headphones and signaled their interpretation of the emotion represented in the passage by pressing a button above pictures of faces representing neutral, sad, angry, and happy emotions. The pictures were accompanied by printed words labeling the emotion represented by the face.
  • The Ps listened to a total of 108 sentences each (54 neutral, 18 sad, 18 angry, and 18 happy.)
  • The experimental task lasted 45 minutes.
  • The investigators analyzed the results using a 2 x 4 ANOVA (group x emotions) and correlations (age and number of correct responses)

 

  1. What were the results of the inferential statistical testing?
  • There was no significant between AS and TYP groups for average comprehension scores.
  • What was the statistical test used to determine significance? ANOVA
  • Were effect sizes provided? No
  • Were confidence interval (CI) provided? No

 

  1. What were the results of the correlational statistical testing?
  • Correlations between discrimination scores and age were not significant.
  • The statistical test used to determine correlation was two- tailed bivariate correlation coefficient.
  1. What were the results of the descriptive analysis
  • The investigators provided a confusion index.

– Rarely confused with other emotions: happy and sad

– Angry confused with: neutral, sad and less frequently happy

– Neutral was confused with: sad

 


Nadig & Shaw (2011)

April 10, 2015

ANALYSIS

Comparison Research

 

 

KEY:   

 

ADOS = Autism Diagnostic Observation Schedule

ASD = autism spectrum disorder

CELF-4 = Clinical Evaluation of Language Fundamental, 4th edition

HFA = High Functioning Autism

P = participant

pmh = Patricia Hargrove, blog developer

SCQ = Social Communication Questionnaire

TYP = typical peer

 

 

SOURCE: Nadig, A., & Shaw, H. (2011). Acoustic and perceptual measurement of expressive prosody in High-Functioning Autism: Increased pitch range and what it means to listeners. Journal of Autism and Developmental Disorders, 42, 499-511.

REVIEWER(S): pmh

DATE: March 31, 2015

ASSIGNED GRADE FOR OVERALL QUALITY: B (The highest possible grade, based on the design of the investigation was B+.)

TAKE AWAY: Findings support the contention that Ps with High Functioning Autism (HFA) produced larger than typical pitch ranges in speech of Ps with although group data revealed that listeners did not perceive the pitch variability of speakers with HFA and their typically developing peers to be significantly different. Other measures (mean pitch and rate) were not significantly different in HFA children/adolescents and their typically developing peers (TYP.) Although there were moderate correlations between perceptual and acoustic measures of mean pitch and speaking rate, the correlation between the acoustic and perceptual measure of pitch change/variability was not significant.

 

  1. What type of evidence was identified?
  • What was the type of evidence? 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 there were groups and participants were not randomly assigned to groups, were members of groups carefully matched? Yes

                                                                    

  1. 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? There were 3 experiments. The numbers for each experiment are listed. In addition, judges were used to rate the speech samples from Experiment 2. These judges will be described below as “raters.”

 

EXPERIMENT 1 AND 2 (the data from the same participants, Ps, were analyzed in Experiments 1 and 2)

  • total # of participant: 28
  • # of groups: 2
  • # of participants in each group: 15, 13
  • List names of groups: High Functioning Autism (HFA) = 15; Typically Developing (TYP) = 13
  • Did all groups maintain membership? Yes

 

EXPERIMENT 3 (Five of the participants, from the High-Functioning Autism group and 10 of the typically developing, TYP, group also participated in Experiments 1 and 2.)

  • total # of participant: 26
  • # of groups:  2
  • # of participants in each group: 15, 11
  • List names of groups: HFA = 15; TYP = 11
  • Did all groups maintain membership? Yes

 

RATERS

  • total # of participant: 32
  • # of groups: 1
  • # of participants in the group: 32
  • List names of group: raters
  • Did all groups maintain membership? Yes
  • The following variables were described:

EXPERIMENT 1 AND 2

  • age: mean age HFA = 11-0 years; TYP = 11-0 years
  • gender: HFA 13m, 2f; TYP 11m, 2f
  • cognitive skills: mean IQ HFA = 105; TYP = 111
  • language: Clinical Evaluation of Language Fundamental, 4th edition (CELF-4) mean HFA = 109; TYP = 115
  • Measures of Autistic Symptoms

– HFA

  • Social Communication Questionnaire (SCQ, parental report) — mean = 26
  • Autism Diagnostic Observation Schedule (ADOS) algorithm score–13
  • ADOS total score (sum of all items) –26

– TYP

  • SCQ (parental report)—2
  • ADOS algorithm score—not applicable (NA)
  • ADOS total score (sum of all items)–NA

EXPERIMENT 3

  • age: mean age HFA = 10-6 years; TYP = 10-08 years
  • gender: HFA 12m, 3f; TYP 9m, 2f
  • cognitive skills: mean IQ HFA = 111; TYP = 116
  • language: CELF-4 mean HFA = 108; TYP = 117

– HFA

  • Social Communication Questionnaire (SCQ, parental report) — mean = 26
  • Autism Diagnostic Observation Score (ADOS) algorithm score–15
  • ADOS total score (sum of all items) –25

– TYP

  • SCQ (parental report)—2
  • ADOS algorithm score—not applicable (NA)
  • ADOS total score (sum of all items)–NA

 

RATERS (from Experiment 2)

  • educational level of rater: Applied Masters students in Communication Sciences and Disorders
  • Were the groups similar before intervention began? Yes. With the exception of the SCQ, the TYP and HFA groups were similar                                                      
  • Were the communication problems adequately described? No
  • disorder type: HFA
  • functional level: performance of HFA group was within normal limits (WNL) on the CELF-4 but the Ps with HFA evidenced social communication problems as noted by their score on the SCQ. All Ps in the HFA group preformed above the 15 on the SCQ which is consistent with the diagnosis of autism spectrum disorder (ASD.) None of the TYP group scored 15 or higher on the SCQ.

 

 

  1. What were the different conditions for this research?
  • Subject (Classification) Groups? Yes–diagnostic classification (HFA; TYP)
  • Experimental Conditions? No
  • Criterion/Descriptive Conditions? Yes

– Experiments 1 and 2: face-to-face conversational speech

– Experiment 3: referential communication task

 

  1. Were the groups controlled acceptably? Yes

 

 

  1. Were dependent measures appropriate and meaningful? Yes

The dependent measures were

EXPERIMENT #1: conversational speech

  • Measure 1: Pitch range
  • Measure 2: Mean Pitch
  • Measure 3: Speech Rate
  • Measure 4: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

EXPERIENT #2: conversational speech

  • Measure 5: Pitch range—used the same data as Experiment 1
  • Measure 6: Mean pitch– used the same data as Experiment 1
  • Measure 7: Speech rate– used the same data as Experiment 1
  • Measure 8: Overall perceptual impression of normalcy
  • Measure 9: Perceptual rating of pitch change
  • Measure 10: Perceptual rating of mean pitch
  • Measure 11: Perceptual rating of speaking rate
  • Measure 12: Relationship between acoustic and perfection of measures
  • Measure 13: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

EXPERIMENT #3: referential communication task

  • Measure 14: Pitch range
  • Measure 15: Mean pitch
  • Measure 16: Speech rate
  • Measure 17: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

The dependent measures that are subjective are

EXPERIMENT #1: conversational speech

  • Measure 4: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

EXPERIENT #2: conversational speech

  • Measure 9: Perceptual rating of pitch change—same data as Experiment #1
  • Measure 10: Perceptual rating of mean pitch—same data as Experiment #1
  • Measure 11: Perceptual rating of speaking rate—same data as Experiment #1
  • Measure 12: Relationship between acoustic and perfection of measures
  • Measure 13: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

EXPERIMENT #3: referential communication task

  • Measure 17: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

 

– The dependent measures that are objective are

EXPERIMENT #1: conversational speech

  • Measure 1: Pitch range
  • Measure 2: Mean Pitch
  • Measure 3: Speech Rate

EXPERIENT #2: conversational speech

  • Measure 5: Pitch range—same data as Experiment #1
  • Measure 6: Mean pitch—same data as Experiment #1
  • Measure 7: Speech rate—same data as Experiment #1

EXPERIMENT #3: referential communication task

  • Measure 14: Pitch range
  • Measure 15: Mean pitch
  • Measure 16: Speech rate

                                         

 

  1. Were reliability measures provided?

– Interobserver for analyzers? No

Intraobserver for analyzers? No

Treatment fidelity for investigators? No

 

 

  1. Description of design:
  • This investigation involved 3 experiments:

— Experiment 1: The acoustical analysis of selected aspects of prosody from brief samples of conversation

— Experiment 2: The perceptual rating of selected aspects of prosody from brief samples of conversation

— Experiment 3: The acoustical analysis of selected aspects of prosody from a referential communication task

  • Participants for each of the investigations were school-age children (8 to 14 years old) who had been diagnosed as HFA and their typically developing peers.
  • The investigators elicited the samples from the Ps and then analyzed them acoustically or perceptually to extract the measures under consideration.

 

 

  1. What were the results of the statistical (inferential) testing?

 

– The comparisons that are significant (p ≤ 0.05) are

EXPERIMENT #1: conversational speech

  • Measure 1: Pitch range—HFA significantly wider than TYP

EXPERIENT #2: conversational speech

  • Measure 5: Pitch range — HFA significantly wider than TYP (used same data as Experiment 1)
  • Measure 8: Overall perceptual impression of normalcy—TYP was significantly higher than HFA

EXPERIMENT #3: referential communication task

  • Measure 14: Pitch range— HFA significantly wider than TYP

– The statistical tests that were used to determine significance were

  • t-test:
  • Mann-Whitney U

– Were effect sizes provided? Yes

EXPERIMENT #1: conversational speech

  • Measure 1: Pitch range – r = 0.67 (moderate effect)
  • Measure 2: Mean Pitch – r = 0.30 (small effect)
  • Measure 3: Speech Rate – r = 0.25 (small effect)

EXPERIENT #2: conversational speech

  • Measure 5: Pitch range – r = 0.67 (moderate effect) same data as Experiment 1
  • Measure 6: Mean pitch – r = 0.30 (small effect) same data as Experiment 1
  • Measure 7: Speech rate – r = 0.25 (small effect) same data as Experiment 1
  • Measure 8: Overall perceptual impression of normalcy – r = 0.48 (small effect)
  • Measure 9: Perceptual rating of pitch change – r = 0.17 (no effect)
  • Measure 10: Perceptual rating of mean pitch – r 0.09 (no effect)
  • Measure 11: Perceptual rating of speaking rate — r 0.22 (small effect)

EXPERIMENT #3: referential communication task

  • Measure 14: Pitch range –r = 0.40 (small effect)
  • Measure 15: Mean pitch –r = 0.22 (small effect)
  • Measure 16: Speech rate — r = 0.03 (no effect)

Were confidence interval (CI) provided? No

Were correlational statistics provided: Yes

– The results of correlational analyses are

EXPERIMENT #1: conversational speech

  • Measure 4: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

— pitch range: none of the correlations were significant for either group (HFA, TYP)

— mean pitch and rate: not reported

EXPERIENT #2: conversational speech

  • Measure 12: Relationship between acoustic and perfection of measures

HFA:

  • pitch change—acoustic and perceptual measures –not significantly correlated (NOTE: both HFA and TYP were not significantly correlated despite the finding that acoustic measures of pitch change were significantly higher for HFA. Visual inspection of the scatterplots suggested that the relationship between acoustic and perceptual measures was more linear in TYP and HFA was flat suggesting different patterns.)
  • mean pitch– acoustic and perceptual measures: significant correlation (r = 0.53, moderate correlation)
  • mean rate — acoustic and perceptual measures: significant correlation (r = 0.65, strong correlation)

 

     – TYP:

  • pitch change—acoustic and perceptual measures –not significantly correlated (NOTE: both HFA and TYP were not significantly correlated despite the finding that acoustic measures of pitch change were significantly higher for HFA. Visual inspection of the scatterplots suggested that the relationship between acoustic and perceptual measures was more linear in TYP and HFA was flat suggesting different patterns.)
  • mean pitch– acoustic and perceptual measures—correlation not significant
  • mean rate — acoustic and perceptual measures: significant correlation (r = 0.87, strong correlation
  • Measure 13: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

no significant correlations for either group

EXPERIMENT #3: referential communication task

  • Measure 17: Relationship between acoustic measures and P characteristics such as IQ, language level, severity of autism, etc.

     – HFA

  • pitch range not significantly correlated with any of the P characteristics
  • correlations with the other acoustic measures were not reported

     – TYP

  • pitch range not significantly correlated with any of the P characteristics

 

  1. Brief summary of clinically relevant results:
  • Using conversational speech and speech during a referential communication task, the investigators determined that Ps with HFA produced speech with larger pitch ranges than typically developing peers. Thus, clinicians (Cs) = should expect a broader pitch range in speech of Ps with HFA rather than a smaller one (which one might expect from the extant clinical literature) using acoustic measures.
  • Using conversational, the investigators determined that raters did not perceive differences between in pitch variability of Ps with HFA and their typically developing peers to be significantly different. Therefore, Cs should not expect that they will perceive the pitch variability to be larger than typically developing peers. If Cs do perceive the P to be monotonal, acoustic measurements may be in order to clarify pitch variability. Despite the fact that there were no significant differences in overall performance, analysis of individual performances suggested that raters judged Ps with HFA to have more extreme (broad and narrow) pitch ranges.
  • The raters judged the speech of Ps with HFA to be significantly more atypical than the TYP group. There is, therefore, evidence for perceived differences between Ps with HFA and their typically developing peers. The cause of this difference is beyond the scope of this paper.

 

ASSIGNED GRADE FOR QUALITY OF EXTERNAL EVIDENCE: B


Shriberg et al. (2001)

December 28, 2014

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

 

Key:

AS = Asperger syndrome

ASD = Autism Spectrum Disorder

HFA = High Functioning Autism

NA = not applicable

P = participant

PEPPER = Programs to Examine Phonetic and Phonologic Evaluation Records

pmh = Patricia Hargrove, blog developer

PVSP = Prosody-Voice Screening Profile

WNL = within normal limits

wpm = words per minute

 

 

SOURCE: Shriberg, L. D., Paul, R., McSweeney, J. L., Klin, A., Cohen, D. J., & Volkmer, F. R. (2001). Speech and prosody characteristics of adolescents and adults with High-Functioning Autism and Asperger syndrome. Journal of Speech, Language, and Hearing Research, 44, 1097-1115.

 

REVIEWER(S): pmh

 

DATE: December 27, 2014

ASSIGNED GRADE FOR OVERALL QUALITY: C+ (The highest possible grade, based on the design of the investigation was C+.)

 

POPULATION: Autism Spectrum Disorders, High Functioning Autism, Asperger Syndrome; Adults, Adolescents

 

PURPOSE: To describe the segmental and nonsegmental aspects of the conversational speech of adolescents and adults with High-Functioning Autism (HFA) and Asperger syndrome (AS). (This review will focus only on the nonsegmental/prosodic results of the investigation.)

 

INSIGHTS ABOUT PROSODY:

  • For the most part, speakers with AS and HFA present with similar prosodic pattern. However, both groups tend to differ from a comparison group of adolescent and adult males who are considered to be typical speakers (TS).
  • The investigators found that the AS and/or HFA groups differed from the TS group on the following prosodic variables:

– Overall Phrasing errors such as

  1. sound repetition errors
  2. word repetition errors
  3. 1 word repetition errors

– Slow/pause time errors (within the Rate category)

– Overall Stress errors

– Excessive/equal/misplaced stress (within the Stress category)

– Too loud (within the Loudness category)

  • The exceptions to the characterizations of similar prosody among speakers with AS and HAS are

– sound syllable repetitions (AS Ps produced significantly more)

– repetition and revision errors (AS Ps produced significantly more)

– slow articulation/pause errors (HFA Ps produced significantly more)

  • Other features:

– The investigators provided a comprehensive and thoughtful review of the literature pertaining to the prosody of speakers who have been diagnosed with Autism Spectrum Disorders (ASD).

 

 

  1. What type of evidence was identified? Retrospective, Nonrandomized Group Comparison Design
  1. Group membership determination:
  2. If there were groups of participants were members of groups matched? Yes
  3. The matching strategy involved
  • All the participants (Ps) were male.
  • There were no significant differences for AS and HFA groups on the following variables:

– age

– average word per utterance

– performance on Intelligence tests, Vineland Adaptive Behavior Scale, and the Test of Language Competence

  • There was not a significant difference between the typical comparison speakers (TS) and the AS and HFA speakers on the following variables:

– age

– number of words processed

  1. Was participants’ communication status concealed?
  2. from participants? No
  3. from assessment administrators? No
  4. from data analyzers? Unclear. Even if the analyzers were not informed of group identify, it is highly likely they would be able to distinguish TS from AS and HFA speakers on the basis of their spontaneous speech.

                                                                    

 

  1. Were the groups/participants adequately described? Yes
  2. How many participants were involved in the study?
  • total # of participants: 83
  • was group membership maintained throughout the experiment? Yes
  • # of groups: 3
  • List names of groups: Asperger syndrome (AS), High-Functioning Autism (HFA), and Typical Speakers (TS)
  • # of participants in each group: AS = 15, HFA = 15, TS = 53

                       

  1. The following variables were controlled and/or described:
  • age: overall range was 10 – 49 years; mean AS age = 20.7; mean HFA age = 21.6; mean TS = 26.4
  • gender: all Ps were male
  • cognitive skills: all within normal limits (WNL);
  • performance on language tests: Composite Score on Test of Language Competence: AS = 90.7; HFA = 88.4; TS not tested
  • adaptive behavior: Composite Score: AS = 58.9; HFA = 55.3; TS not tested
  • performance on Autism Diagnostic Observation Scale –Communication Subtest: AS = 3.7; HFA = 5.4; TS not tested (AS and HFA significantly different)
  • performance on Autism Diagnostic Observation Scale –Socialization Subtest: AS = 9.9; HFA = 12.0; TS not tested (AS and HFA significantly different)

 

  1. Were the communication problems adequately described? Yes
  • disorder type: (List) ASD, social language

 

  1. What were the different conditions for this research?
  2. Subject (Classification) Groups?

Yes: AS, HFA, TS

                                                               

  1. Experimental Conditions?

 

  1. Criterion/Descriptive Conditions? Yes: Performance on the Prosody-Voice Screening Profile (PVSP) and Programs to Examine Phonetic and Phonologic Evaluation Records (PEPPER)

 

  1. Were the groups controlled acceptably? Yes

 

 

  1. Were dependent measures appropriate and meaningful? Yes

 

  1. The Prosody-Related Outcomes:
  • Dependent Measure #1: Percentage of utterances coded as prosodically appropriate on the PVS-
  • Dependent Measure #2: Percentage of appropriate phrasing on the PVSP
  • Dependent Measure #3: Percentage of sound/syllable repetition errors on the PVSP
  • Dependent Measure #4: Percentage of word repetition errors on the PVSP
  • Dependent Measure #5: Percentage of sound/syllable and word repetition errors on the PVSP
  • Dependent Measure #6: Percentage of more than one repetition errors on the PVSP
  • Dependent Measure #7: Percentage of sound/syllable repetition errors on the PVSP
  • Dependent Measure #8: Percentage of one-word revisions on the PVSP
  • Dependent Measure #9: Percentage of more than one word revisions on the PVSP
  • Dependent Measure #10: Percentage of repetition and revision errors on the PVSP

 

  • Dependent Measure #11: Percentage of appropriate rate on the PVSP
  • Dependent Measure #12: Percentage of slow articulation/pause errors on the PVSP
  • Dependent Measure #13: Percentage of slow/pause time errors on the PVSP
  • Dependent Measure #14: Percentage of fast errors on the PVSP
  • Dependent Measure #15: Percentage of fast/acceleration errors on the PVSP
  • Dependent Measure #16: Percentage of appropriate stress on the PVSP
  • Dependent Measure #17: Percentage of multisyllabic word stress errors on the PVSP
  • Dependent Measure #18: Percentage of reduced/equal stress errors on the PVSP
  • Dependent Measure #19: Percentage of excessive/equal/misplaced errors on the PVSP
  • Dependent Measure #20: Percentage of multiple stress errors on the PVSP
  • Dependent Measure #21: Percentage of appropriate loudness on the PVSP
  • Dependent Measure #22: Percentage of too soft errors on the PVSP
  • Dependent Measure #23: Percentage of too loud errors on the PVSP
  • Dependent Measure #24: Percentage of appropriate pitch on the PVSP
  • Dependent Measure #25: Percentage of low pitch/glottal fry errors on the PVSP
  • Dependent Measure #26: Percentage of low pitch errors on the PVSP
  • Dependent Measure #27: Percentage of high pitch/falsetto errors on the PVSP
  • Dependent Measure #28: Percentage of high pitch errors on the PVSP
  • Dependent Measure #29: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Phrasing
  • Dependent Measure #30: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Rate
  • Dependent Measure #31: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Stress
  • Dependent Measure #32: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Loudness
  • Dependent Measure #33: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Pitch
  • Dependent Measure #34: Words per minute (volubility)
  1. NON-PROSODIC OUTCOMES (these outcomes will not be analyzed or summarized in this review):
  • Dependent Measure #35: Percentage of appropriate laryngeal quality on the PVSP
  • Dependent Measure #36: Percentage of breathiness errors on the PVSP
  • Dependent Measure #37: Percentage of roughness errors on the PVSP
  • Dependent Measure #38: Percentage of strained errors on the PVSP
  • Dependent Measure #39: Percentage of break/shift/tremulous errors on the PVSP
  • Dependent Measure #40: Percentage of register errors on the PVSP
  • Dependent Measure #41: Percentage of diplophonia on the PVSP
  • Dependent Measure #42: Percentage of multiple laryngeal feature errors on the PVSP
  • Dependent Measure #43: Percentage of appropriate resonance quality on the PVSP
  • Dependent Measure #44: Percentage of nasality errors on the PVSP
  • Dependent Measure #45: Percentage of denasality errors on the PVSP
  • Dependent Measure #46: Percentage of nasopharyngeal errors on the PVSP
  • Dependent Measure #47: Percentage of excluded utterances on the PVSP
  • Dependent Measure #48: Percentage of utterances excluded due to context issues (issue types were also compared)
  • Dependent Measure #49: Percentage of utterances excluded due to environmental issues (issue types were also compared)
  • Dependent Measure #50: Percentage of utterances excluded due to register issues (issue types were also compared)
  • Dependent Measure #51: Percentage of utterances excluded due to state issues (issue types were also compared)
  • Dependent Measure #52: Percentage Consonants Correct
  • Dependent Measure #53: Percentage Vowels/Diphthongs Correct
  • Dependent Measure #54: Percentage of Phonemes Correct
  • Dependent Measure #55: Percentage of Consonants Correct –Revised
  • Dependent Measure #56: Percentage Vowels/Diphthongs Correct–Revised
  • Dependent Measure #57: Intelligibility Index
  • Dependent Measure #58: Percentage of Ps with residual distortions
  • Dependent Measure #59: Type of residual distortion errors
  • Dependent Measure #60: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Laryngeal Quality
  • Dependent Measure #61: Percentage of HFA and AS Ps with more than 20% of their utterance coded as a fail on the PVSP for Resonance Quality

 

  1. All the dependent measures were subjective.

 

  1. None of the the dependent/ outcome measures were objective.

                                         

 

  1. Were reliability measures provided?
  2. Interobserver for analyzers? Yes. Interobserver reliability for PVSP coding ranged from 14.3% (stress coding) to 95.8%

 

  1. Intraobserver for analyzers? Yes. Intraobserver reliability for PVSP coding ranged from 46.7% (stress coding) to 94.5%

 

  1. Treatment fidelity for investigators? Not Applicable

 

 

  1. Description of design:
  • The investigators coded preexisting spontaneous language samples using the PVSP.
  • They compared the performances of the 2 clinical groups (HFA, AS) and the comparison group (TS) using nonparametric inferential statistics.
  • The investigators also performed correlational analyses among cognitive, linguistic, and adaptive behavior functioning and prosody measures. (Because the correlations were generally nonsignificant and low to moderate, they received only limited attention in the paper and will not be discussed in the review.)

 

  1. What were the results of the inferential statistical testing?
  2. The comparisons that are significant are bolded; the number in parentheses represents the page number of the article where the result can be found.
  • Dependent Measure #1: Percentage of utterances coded as prosodically appropriate on the PVSP– :Percentage of appropriate utterances on the PVSP– Ps AS and Ps with HFA had significantly lower total percentages of appropriate responses in the PVSP than TS (p. 1106)

 

  • Dependent Measure #2: Percentage of appropriate phrasing on the PVSP– Ps AS and Ps with HFA had significantly lower percentages of appropriate phrasing than TS (p. 1106)
  • Dependent Measure #3: Percentage of sound/syllable repetition errors on the PVSP—HFA Ps had significantly more utterances coded for this error than TS Ps (p. 1106); AS Ps had significantly more utterances coded for this error than HFA Ps (p. 1106)
  • Dependent Measure #4: Percentage of word repetition errors on the PVSP– HFA Ps had significantly higher more utterances coded for this error than TS Ps (p. 1106)
  • Dependent Measure #5: Percentage of sound/syllable and word repetition errors on the PVSP
  • Dependent Measure #6: Percentage of more than one repetition errors on the PVSP
  • Dependent Measure #7: Percentage of sound/syllable repetition errors on the PVSP
  • Dependent Measure #8: Percentage of one word revisions on the PVSP– HFA Ps had significantly higher more utterances coded for this error than TS Ps (p. 1106)
  • Dependent Measure #9: Percentage of more than one word revisions on the PVSP
  • Dependent Measure #10: Percentage of repetition and revision errors on the PVSP– AS Ps had significantly more utterances coded for this error than HFA Ps (p. 1106)
  • Dependent Measure #11: Percentage of appropriate rate on the PVSP –overall all 3 groups (HFA, AS, TS) had over 90% of their utterances coded for appropriate rate.

 

  • Dependent Measure #12: Percentage of slow articulation/pause errors on the PVSP—Ps with HFA had significantly more utterances coded for this error compared with AS and with TS (p. 1007)

 

  • Dependent Measure #13: Percentage of slow/pause time errors on the PVSP Ps with HFA had significantly more utterances coded for this error compared with TS (p. 1007)

 

  • Dependent Measure #14: Percentage of fast errors on the PVSP
  • Dependent Measure #15: Percentage of fast/acceleration errors on the PVSP
  • Dependent Measure #16: Percentage of appropriate stress on the PVSP–significantly more Ps with HFA and AS had inappropriate stress than TS (p. 1108); the 3 groups (HFA, AS, TS) differed significantly in the number of utterances codes as containing appropriate stress (p. 1107)

 

  • Dependent Measure #17: Percentage of multisyllabic word stress errors on the PVSP
  • Dependent Measure #18: Percentage of reduced/equal stress errors on the PVSP
  • Dependent Measure #19: Percentage of excessive/equal/misplaced errors on the PVSP– Ps with HFA and AS had significantly more utterances coded for this error compared with TS Ps (p. 1007)
  • Dependent Measure #20: Percentage of multiple stress errors on the PVSP
  • Dependent Measure #21: Percentage of appropriate loudness on the PVSP –overall more than 90% of utterances for all 3 groups (AS, HFA, TS) were coded as appropriate (p. 1107)
  • Dependent Measure #22: Percentage of too soft errors on the PVSP
  • Dependent Measure #23: Percentage of too loud errors on the PVSP – AS and HFA Ps were more likely to be coded with this error than TS Ps )
  • Dependent Measure #24: Percentage of appropriate pitch on the PVSP– overall more than 90% of utterances for all 3 groups (AS, HFA, TS) were coded as appropriate (p. 1107)
  • Dependent Measure #25: Percentage of low pitch/glottal fry errors on the PVSP
  • Dependent Measure #26: Percentage of low pitch errors on the PVSP
  • Dependent Measure #27: Percentage of high pitch/falsetto errors on the PVSP
  • Dependent Measure #28: Percentage of high pitch errors on the PVSP
  • Dependent Measure #29: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Phrasing– significantly more AS Ps produced utterances with inappropriate stress than TS Ps (p. 1108)
  • Dependent Measure #30: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Rate

 

  • Dependent Measure #31: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Stress—significantly more HFA and AS Ps produced utterances with inappropriate stress than TS Ps (p. 1108)
  • Dependent Measure #32: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Loudness
  • Dependent Measure #33: Percentage of HFA and AS Ps with more than 20% of their utterances coded as a fail on the PVSP for Pitch
  • Dependent Measure #34: Words per minute–WPM (volubility) –Ps with AS produced significantly more WPM than Ps with HFA. (Investigators did not analyze this measure for TS Ps.)
  1. What were the statistical tests used to determine significance? List the outcome number after the appropriate statistical test: Kruskal-Wallis, Wilcoxon-Mann-Whitney, Tests of Proportion
  1. Were effect sizes provided?   No
  1. Were confidence interval (CI) provided? No

 

 

  1. What were the results of the correlational statistical testing? Some correlational testing was reported but it was not a major focus of the investigation and will not be reviewed here.
  1. What were the results of the descriptive analysis?

For Dependent Measure #19: Percentage of excessive/equal/misplaced errors on the PVSP–

The following results were not subjected to inferential testing but the investigators described differences in AS and HFA groups pertaining to the types of errors in the excessive/equal/misplaced stress subcode:

– Ps with HFA (39% of all codes for this error type) were more likely to produce word stress errors than AS Ps (22% of all codes for this error type)

– Ps with AS (54% of all codes for this error type) were more likely to produce prolongation errors than HA Ps (32% of all codes for this error type)

– Both HFA (25% of all codes for this error type) and AS (20% of all codes for this error type) were relatively unlikely to produce blocks (p. 1107)

 


Trauner (2008)

December 13, 2014

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

Key:

NA = not applicable

pmh = Patricia Hargrove, blog developer

RHBD = right hemisphere brain damage

 

SOURCE: Trauner, D. A. (2008). Right hemisphere brain damage in children. Perspectives on Neurophysiology and Neurogenic Speech and Language Disorders, 18, 73-81.

 

REVIEWER(S): pmh

 

DATE: December 14, 2014

ASSIGNED GRADE FOR OVERALL QUALITY: D – (The highest possible grade for this review was D due to the nature of the design.)

 

POPULATION: Right Hemisphere Brain Damage; Children

 

PURPOSE: To describe the cognitive and linguistic functions affected by right hemisphere brain damage (RHBD) in children

 

INSIGHTS ABOUT PROSODY:

Although the author addresses numerous cognitive and linguistic functions (e.g., intellectual test performance, language comprehension, expressive language measures, facial recognition, visual spatial skills) that are affected by RHBD in children, only prosody will be discussed in this review. Only 2 investigations concerned with prosody were reviewed. Together they suggest that children with RHBD are likely to experience problems with the production of linguistic prosody as well as the comprehension and production of affective prosody.

 

 

  1. What type of evidence was identified? Secondary Research

What type of secondary review? Narrative Review

 

  1. Were the results valid? Yes
  2. Was the review based on a clinically sound clinical question? Yes
  3. Did the reviewer clearly describe reasonable criteria for inclusion and exclusion of literature in the review (i.e., sources)? No, the author did not describe criteria.
  4. Author noted that she reviewed the following resources: The author did not describe the search strategy.
  5. Did the sources involve only English language publications? Yes
  6. Did the sources include unpublished studies? Yes
  7. Was the time frame for the publication of the sources sufficient? Yes
  8. Did the reviewer identify the level of evidence of the sources? No
  9. Did the reviewers describe procedures used to evaluate the validity of each of the sources? No
  10. Was there evidence that a specific, predetermined strategy was used to evaluate the sources? No
  11. Did the reviewer or review teams rate the sources independently? No
  12. Were interrater reliability data provided? NA
  13. If the reviewers provided interrater reliability data, list them: NA
  14. If there were no interrater reliability data, was an alternate means to insure reliability described? NA
  15. Were assessments of sources sufficiently reliable? NA
  16. Was the information provided sufficient for the reader to undertake a replication? No
  17. Did the sources that were evaluated involve a sufficient number of participants? Unclear
  18. Were there a sufficient number of sources? No but that is the status of the literature.
  1. Description of outcome measures:

NOTE: Only behaviors concerned with prosody as an outcome or as a dependent variable will be described here.

  • Outcome #1: Expression of affective and linguistic prosody (Trauner et al., 1996)
  • Outcome #2: Receptive prosody (Cohen et al., 1994)

 

 

  1. Description of results:
  2. What evidence-based practice (EBP) measures were used to represent the magnitude of the treatment/effect size?  NA
  3. Summarize overall findings of the secondary review:
  • Outcome #1: Expression and comprehension of affective and linguistic prosody (Trauner et al., 1996)

Trauner et al. (1996) reported that children with early focal lesions of the right or left hemisphere had difficulty with the production of linguistic and affective prosody and only those with RHBD had trouble with the comprehension of affective prosody.

  • Outcome #2: Receptive prosody (Cohen et al., 1994)

Cohen et al. (1994) reported that neuropsychological testing revealed that children with RHBD had trouble with overall receptive prosody. Children with left hemisphere brain damage did not evidence similar problems.

  • Overall,

– children with RHBD are likely to experience problems with

  1. the comprehension and production of affective prosody
  2. the production of linguistic prosody
  3. performance on neuropsychological measures of overall receptive prosody

–   children with left hemisphere brain damage seem only to be challenged by

  1. the production of (linguistic and affective) prosody .

 

  1. Were the results precise? Unclear, this information was not provided by the author.
  2. If confidence intervals were provided in the sources, did the reviewers consider whether evaluations would have varied if the “true” value of metrics were at the upper or lower boundary of the confidence interval? NA
  3. Were the results of individual studies clearly displayed/presented? Yes, but the author only provided general statements about the findings.
  4. For the most part, were the results similar from source to source? Yes
  5. Were the results in the same direction? Yes
  6. Did a forest plot indicate homogeneity? NA
  7. Was heterogeneity of results explored? No
  8. Were the findings reasonable in view of the current literature? Yes
  9. Were negative outcomes noted? No

 


Penner et al. (2001)

October 29, 2014

NATURE OF PROSODIC DISORDERS

ANALYSIS FORM

Key:

Fo = fundamental frequency

P = participant or patient

PD = Parkinson’s disease

pmh = Patricia Hargrove, blog developer

TA = typical adult

 

SOURCE: Penner, H., Miller, N, Hertrich, I., Ackermann, H., & Schumms, F. (2001). Dysprosody in Parkinson’s disease: An investigation of intonation patterns. Clinical Linguistics and Phonetics, 15, 551-566.

 

REVIEWER(S): pmh

 

DATE: October 25, 2014

ASSIGNED GRADE FOR OVERALL QUALITY: C+ (Based on the design, the highest possible grade was B+.)

 

POPULATION: Parkinson’s disease (PD); Adults

 

PURPOSE: To describe the interaction of accent patterns, accent timing, and medication on the peak height of fundamental frequency (Fo) in German speakers with Parkinson’s disease (PD).

 

INSIGHTS ABOUT PROSODY:

  • Compared to typical adults (TA), German speakers with PD tended to
  1. produce reduced Fo at the peak of the intonation contour and
  2. locate Fo peak on the targeted accented syllable more frequently. (This second factor is noted in the abstract but I could not find it in the results.)
  • Medication had variable effects on the Fo peaks of Ps with PD.

 

  1. What type of evidence was identified?
  • Combined Design– Single Subject Design with Specific Participants: ABC design and Prospective, Nonrandomized Group Comparison Design
  1. Group membership determination:
  2. If there were groups of participants were members of groups matched? Yes
  3. The matching strategy involved the following:
  • The TA group was matched to PD group by age and status of being a native, monolingual speaker of German. The TA group had no history of neurological impairments and were hospitalized due to orthopedic issues.
  1. Was participants’ communication status concealed?
  2. from participants? No
  3. from assessment administrators? No
  4. from data analyzers? No

                                                                    

 

  1. Were the groups/participants adequately described? Yes
  2. How many participants were involved in the study?
  • total # of participants: 6
  • was group membership maintained throughout the experiment? Yes
  • # of groups: 2
  • List names of groups: Speakers with Parkinson’s disease (PD) and speakers with no known neurological problems—typical adults (TA)
  • # of participants in each group: PD = 3Ps; TA = 3Ps

           

  1. The following variables were actively controlled described
  • age: 72-74 years
  • gender: all Ps with PD were female
  • cognitive skills: all Ps with PD- no dementia
  • depression: none of Ps with PD had evidence of severe depression at time of investigation
  • language: all Ps were monolingual, native speakers of German
  • diagnosis: all Ps with PD had been diagnosed with idiopathic Parkinson’s disease
  • severity of PD: on the Hoehn & Yahr Scale—III to IV; on the Webster Scale –14 to 17
  • duration of PD: 9 to 14 years
  • neurological problems prior to the diagnosis of PD: none for all Ps with PD
  • medication: all Ps with PD received levodopa

 

  1. Were the communication problems adequately described? No
  • disorder type: dysarthria associated with PD

 

 

  1. What were the different conditions for this research?
  2. Subject (Classification) Groups? Yes: PD and TA
  3. Experimental Conditions? Yes
  • Accent patterns (4: Accent 1, Accent 2, Accent 3, Question rise)
  • Affect (2: ‘really angry’, neutral question/statement)
  • Medication level (2: Low, High)

    – Low medication = no medication and first 2 recording after receiving medications

– High medication = 3rd, 4th, and 5th recordings after receiving medications

 

  1. Were the groups controlled acceptably? Unclear. For example, it is not clear if TA Ps were females and the groups produced an unequal number sentences.

 

 

  1. Were dependent measures appropriate and meaningful? Yes
  2. The dependent measures were
  • Dependent Measure #1: Relative rise of Fo on accent 1 of declarative sentences
  • Dependent Measure #2: Relative rise of Fo on accent 2
  • Dependent Measure #3: Relative rise of Fo on accent 3
  • Dependent Measure #4: Relative rise of Fo on question rises
  • Dependent Measure #5: Distribution of accent pattern: peak, valley, fall, incomplete linking, linking
  • Dependent Measure #6: Percentage of middle timing in peaks and linking patterns for accent 1
  • Dependent Measure #7: Percentage of late timing in peaks and linking patterns for accent 1
  • Dependent Measure #8: Influence of medication on Fo peak
  1. The dependent measure that was subjective is
  • Dependent Measure #5: Distribution of accent pattern: peak, valley, fall, incomplete linking, linking

 

  1. The dependent measures that were objective are
  • Dependent Measure #1: Relative rise of Fo on accent 1 of declarative sentences
  • Dependent Measure #2: Relative rise of Fo on accent 2
  • Dependent Measure #3: Relative rise of Fo on accent 23
  • Dependent Measure #4: Relative rise of Fo on question rises
  • Dependent Measure #6: Percentage of middle timing in peaks and linking patterns for accent 1
  • Dependent Measure #7: Percentage of late timing in peaks and linking patterns for accent 1
  • Dependent Measure #8: Influence of medication on Fo peak

 

                                         

  1. Were reliability measures provided?
  2. Interobserver for analyzers? No
  3. Intraobserver for analyzers? No
  4. Treatment fidelity for investigators?

 

 

  1. Description of design:
  • PD and TA Ps produced similar sentences but there were different procedures for the 2 groups.
  • Both groups were recorded producing 12 difference sentences.

– The sentences were in German and each sentence had 3 accented words: accent 1, accent 2, accent 3.

– P were directed to produce the first 6 sentences as declaratives and then repeat the same 6 sentences as questions.

– P was then directed to produce the next 6 sentences as if he/she was “really angry.”

  • The PD group produced recording 5 times a day for 4 consecutive days. Sessions were separated by ½ hour.

– For 4 of the 5 days, PD Ps were administered their scheduled dose of levodopa.

– On a randomly assigned day each PD P did not receive levodopa.

– Low medication = no medication and first 2 recordings after receiving medications

– High medication = 3rd, 4th, and 5th recordings after receiving medications

– Overall, the PD Ps produced 360 sentences.

  • The TA Ps produced examples of the same sentences as the PD group but they only produced 90 sentences.
  • This investigation involved a variety of comparisons: some between group comparisons (descriptive comparisons between TA and PD), some within P comparisons (inferential statistical analysis comparing differ contexts of individual P performance), some descriptions of individual Ps.

 

  1. What were the results of the inferential statistical testing?
  2. Only comparisons that are significant are listed
  • Dependent Measure #8: Influence of medication on Fo peak

– P1 with PD = = with medication Fo of peak height was significantly higher than without medication for accent 1 (p = 0.01); with medication Fo of peak height was significantly (0.05) lower than without medication for accent 2

– P2 with PD = with medication Fo of peak height was significantly higher than without medication for accent 1 (p = 0.005) and for question rise (p = 0.0005)

  1. What was the statistical test used to determine significance? t-test: Satterthwaite adjustment (Only Dependent Measure #8 was tested with inferential statistics.)
  1. Were effect sizes provided? No
  1. Were confidence interval (CI) provided? No

 

 

  1. What were the results of the correlational statistical testing? Correlational statistics were not provided.

 

  1. What were the results of the descriptive analysis
  • Dependent Measure #1: Relative rise of Fo on accent 1

   – comparing Ps with PD and TA controls: Ps with PD tend to produce lower relative rises in F0

  • Dependent Measure #2: Relative rise of Fo on accent 2

   – comparing Ps with PD and TA controls: Ps with PD tend to produce lower relative rises in F0

 

  • Dependent Measure #3: Relative rise of Fo on accent 3

these data were not reported

 

  • Dependent Measure #4: Relative rise of Fo on question rises

   – comparing Ps with PD and TA controls: Overall, Ps with PD tend to produce lower relative rises in F0. However, P3 with PD did produce a higher mean than TA peers.

 

  • Dependent Measure #5: Distribution of accent pattern: peak, valley, fall, incomplete linking, linking

– The investigators did not identify overall patterns that differentiated Ps with PD and their TA peers.

 

  • Dependent Measure #6: Percentage of middle timing in peaks and linking patterns for accent 1

– The investigators defined middle timing as the timing of Fo peaks on the appropriate syllable.

–   P2 and P3 with PD preferred middle timing.

  • Dependent Measure #7: Percentage of late timing in peaks and linking patterns for accent 1

– The investigators defined late timing as the timing of Fo peaks on syllable following the appropriate syllable.

– P1 with PD preferred late timing.