2/6/16 Linguistic Mazes in Children with Disclosures Autism - - PDF document

SLIDE 1

2/6/16 1

Linguistic Mazes in Children with Autism Spectrum Disorder or Attention De:icit Hyperactivity Disorder

LIZBETH H. FINESTACK, PH.D., CCC-SLP KATHERINE BANGERT, DOCTORAL STUDENT, M.A., CCC-SLP SOMER L. BISHOP, PH.D., UNIVERSITY OF CALIFORNIA, SAN FRANCISCO

Disclosures

u We have no relevant financial or nonfinancial relaLonships related

to any informaLon presented.

What are linguistic mazes?

Dollaghan & Campbell (1992) Maze Taxonomy

What are linguistic mazes?

Pauses RepeLLons Revisions Orphans

Filled pause: “I (uh) went shopping” Silent pause: “I… went shopping.” Pauses “We (we) went to the store.” RepeLLons “My brother, I mean my sister said I could go.” Revisions “I went to the… Tomorrow is Friday.” Orphans

Dollaghan & Campbell (1992) Maze Taxonomy

What are linguistic mazes? Why do linguistic mazes occur?

Maze producLon associated with weaknesses in the language and execuLve funcLoning cogniLve domains.

SLIDE 2

2/6/16 2

Maze Framework proposed by Rispoli and colleagues (2001, 2003, & 2008)

Stalls

• Reflect “glitches”

in language encoding

• Include filler

pauses and repeLLons Revisions

• Reflect external

monitoring

• Increase with use
• f more complex

language

Mazes and Speci:ic Language Impairment

• Children with SLI generally produce mazes at a higher rate

than typically developing children

• Dollagahn & Campbell, 1992; Thordardo7r & Weismer, 2002; Navarro-Ruiz

& Rallo-Fabra, 2001; Guo, Tomblin, & Samelson, 2008; NeFelbladt & Hannson, 1999; Wetherell, D., Bo7ng & ConK-Ramsden, 2007

• Thordardobr and Ellis Weismer (2002): Compared to

typically developing children, children with SLI produced:

• Significantly more repeLLons, revisions, and orphans
• BUT fewer filled pauses

What about other special populations? Attention De:icit Hyperactivity Disorder (ADHD)

u Redmond (2004) u Children with ADHD had significantly higher use of mazes

than typically developing children and children with SLI.

u The fluency difficulLes were perhaps caused by limitaLons

in execuLve funcLon in these children.

u Redmond suggested that maze use might be used as a

diagnosLc indicator of ADHD.

Autism Spectrum Disorder

u Shriberg et al. (2006) u Compared adolescent and adult males with high funcLoning

auLsm, Asperger’s syndrome, and typically developing peers

u Individuals with high funcLoning auLsm and those with

Asperger’s syndrome produced more word repeLLons and revisions than peers with typical development

Current Study

u To evaluate Rispoli’s and Thordardobr and Ellis Weismer’s

frameworks, we compared the relaLonships between the producLon of filler pauses, repeLLons, and revisions and cogniLve and language development in children with:

u ADHD u ASD

SLIDE 3

2/6/16 3

Research Questions

• 1. Are there differences in maze use

by children with ADHD and children with ASD?

Predic1on:

• We know in both groups, there are subgroups

with language and execuLve funcLoning weaknesses and both groups have elevated maze use.

• These groups have not been directly

compared before, therefore we did not have a clear predicLon regarding group differences.

• 2. Is maze use context dependent?

Predic1on:

• Recalling sentences requires less encoding

because parLcipants are provided with a model sentence.

• We expected more mazes in conversaLonal

contexts versus recalling sentences tasks.

Research Questions

• 3. Are maze types produced at the

same rate?

Predic1on: Based Thordardobr and Ellis Weismer’s framework which characterizes filler pause use as a pragmaLc funcLon to hold the listener’s akenLon, we expected for:

• The ASD group, fewer filled pauses than

repeLLons and revisions.

• The ADHD group, no significant differences

across maze types.

• 4. How does maze use relate to

language ability?

Predic1on: The Rispoli framework associates filler pauses and repeLLons with encoding difficulty, and revisions with language monitoring ability. Therefore we expected:

• Filler pauses and repeLLons to have a weak

associaLon with language ability.

• Revision use to be posiLvely correlated with

language ability.

Method

u ParLcipants u 16 children with ADHD u 8 children with ASD u Age range: 5-13 years

Participant Characteristics

ASD ADHD Total Mean (min-max) SD Mean (min-max) SD Mean (min-max) SD Age NVIQ VIQ MLU CELF 102.5 70-140 110.25 89-136 111 96-139 5.32 3.5-6.8 61.93 45-78 24.8 16.69 14.04 1.18 10.5 117 70-152 96 75-117 94.88 65-118 4.92 3.2-6.8 59.88 37-82 22.03 11.94 16.61 .90 15.06 112.17 70-152 100.75 75-136 100.25 65-139 5.05 3.2-6.8 60.46 37-82 23.5 15.02 17.32 .996 13.5

Maze Sampling Contexts

Au1sm Diagnos1c Observa1on Schedule

(ADOS; Lord, Ruker, DiLavore, & Risi, 2002) u Consists of four modules that provide social-communicaLve sequences through various structured and unstructured situaLons

Clinical Evalua1on of Language Fundamentals-4: Recalling Sentences Subtest

(CELF-4; Semel, Wiig, Secord, 2003) u Evaluates the child’s ability to listen and repeat verbaLm sentences of increasing length and syntacLc

Transcription and Maze Coding

u Trained research assistants transcribed video

recordings of ADOS and audio recordings of the Recalling Sentences subtest of the CELF

u Transcribed using SystemaLc Analysis of

Language TranscripLons convenLons, including maze variables

u Key study variables included: u mean length of ukerance in morphemes (MLU) u rate of maze types per ukerance:

u filler pauses u repeLLons u revisions

SLIDE 4

2/6/16 4

Transcription and Maze Coding

u Percent agreement for conversaLon:

u RepeLLons 93% u Filled pauses 90% u Revisions 78% u Mazes overall 88%

ADOS Reliability Transcription and Maze Coding

u Percent agreement for Sentence

RepeLLon:

u RepeLLons 78% u Filled pauses 92% u Revisions 89% u Mazes overall 90%

CELF Reliability

Results

• 1. Are there differences in maze use by

children with ADHD and children with ASD?

ADOS: Nonparametric Mann –Whitney U Tests; Bars = +/- 1 SD

0.05 0.1 0.15 Filler Pauses RepeLLons Revisions Rate per UPerance ADHD ASD 0.05 0.1 0.15 0.2 0.25 0.3 0.35 Total Mazes Rate per UPerance p = .49 p = .08 p = .79 p = .38

• 1. Are there differences in maze use by

children with ADHD and children with ASD?

**controlling for utterance length**

ADOS: Nonparametric Mann –Whitney U Tests; Bars = +/- 1 SD

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 11 Word Ukerances 12 Word Ukerances 13 Word Ukerances Percent of UPerances with Mazes ADHD ASD p = .14 p = .68 p = .32

• 1. Are there differences in maze use by

children with ADHD and children with ASD?

CELF: Nonparametric Mann –Whitney U Tests; Bars = +/- 1 SD

0.05 0.1 0.15 0.2 0.25 0.3 0.35 Filler Pauses RepeLLons Revisions Rate per UPerance ADHD ASD 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Total Mazes Rate per UPerance p = .83 p = .61 p = .21 p = .93

SLIDE 5

2/6/16 5

• 1. Are there differences in maze use by

children with ADHD and children with ASD?

**controlling for utterance length**

CELF: Nonparametric Mann –Whitney U Tests; Bars = +/- 1 SD

0.1 0.2 0.3 0.4 0.5 0.6 0.7 11 Word Ukerances 12 Word Ukerances 13 Word Ukerances Percent of UPerances with Mazes ADHD ASD p = .88 p = .53 p = 1.00

• 2. Is maze use context dependent?

ADHD: Related Samples Signed Rank Tests; Bars = +/- 1 SD

0.05 0.1 0.15 0.2 0.25 0.3 0.35 Filler Pauses RepeLLons Revisions Rate per UPerance ADOS CELF 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Total Mazes Rate per UPerance p = .44 p < .01 p < .01 p < .01

• 2. Is maze use context dependent?

**controlling for utterance length**

ADHD: Related Samples Signed Rank Tests; Bars = +/- 1 SD

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 11 Word Ukerances 12 Word Ukerances 13 Word Ukerances Percent of UPerances with Mazes ADOS CELF p = .50 p = .06 p = .50

• 2. Is maze use context dependent?

ASD: Related Samples Signed Rank Tests; Bars = +/- 1 SD

0.05 0.1 0.15 0.2 0.25 0.3 0.35 Filler Pauses RepeLLons Revisions Rate per UPerance ADOS CELF 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Total Mazes Rate per UPerance p = .40 p = .07 p = .33 p = .04

• 2. Is maze use context dependent?

**controlling for utterance length**

ASD: Related Samples Signed Rank Tests; Bars = +/- 1 SD

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 11 Word Ukerances 12 Word Ukerances 13 Word Ukerances Percent of UPerances with Mazes ADOS CELF p = .07 p = .09 p = .02

• 3. Are maze types produced at the same

rate?

0.01 0.02 0.03 Filler Pauses RepeLLons Revisions

ADHD

0.01 0.02 0.03 Filler Pauses RepeLLons Revisions

ASD

ADOS: Related Samples Wilcoxon Signed Rank Tests; Bars = +/- 1 SD

p = .07 P < .01 p = .06

SLIDE 6

2/6/16 6

• 3. Are maze types produced at the same

rate?

0.1 0.2 0.3 0.4 Filler Pauses RepeLLons Revisions

ADHD

0.1 0.2 0.3 0.4 Filler Pauses RepeLLons Revisions

ASD

CELF: Related Samples Wilcoxon Signed Rank Tests; Bars = +/- 1 SD

p = .06 p = .02 p = .04

• 4. How does maze use relate to language ability?

ADOS

ADOS

Age NVIQ VIQ CELF Raw Score MLU Filler Pause Rate Repe11on Rate Revision Rate Total Maze Rate Age NVIQ

• .007

VIQ

• .087

.716** CELF Raw Score .476* .387 .440* MLU .435* .396 .354 .561** Filler Pause Rate

• .268
• .006

.380 .005 .133 Repe11on Rate

• .325

.485* .526** .374 .338 .106 Revision Raze 0.65 .318 .314 .180 .597** .233 .387 Total Maze Rate

• .210

.318 .559** .279 .544* .537** .724** .831**

Summary of Findings

• 1. Are there differences in maze use

by children with ADHD and children with ASD?

• No, no group differences based on ADOS
• r CELF, even when controlling for

sentence length

• Only trend: ASD group more maze

repeLLons than ADHD based on ADOS sample (p = .06; d = .42)

• 2. Is maze use context dependent?
• Yes, for ASD: When controlling for

ukerance length, ASD group tended to produce more mazes based on ADOS (ps = .02-.09; ds = .70-.99)

Summary of Findings

• 3. Are maze types produced at the

same rate?

• No, for ADHD: Produce more revisions

than filler pauses and repeLLons (ps < . 01-.06; ds = .55-1.33)

• For ASD, effect is not robust
• 4. How does maze use relate to

language ability?

• ADOS mazes posiLvely correlated with

NVIQ and MLU, but not CELF raw score

• CELF mazes strongly correlated with age,

but not with VIQ or CELF raw score

Conclusions

u Maze producLon does not appear to

clinically differenLate children with ADHD and those with ASD

u Greater maze use in ADOS context suggests

that mazes reflect encoding weaknesses

u Differences in types of mazes produced likely

reflects difference in underlying cause of revisions and other maze types

u Some suggesLon that maze use strongly

related to language ability based on strong maze and language correlaLons in ADOS sample

Acknowledgements

u We are grateful to all the children and their families who made

this project possible.

u Study compleLon and preparaLon of this manuscript were

supported by a Grant-in-Aid Award from the University of Minnesota and R01HD065277 and R03DC011365 grant awards from the NaLonal InsLtutes of Health.