Signals of CNS damage in the discourse behavior of school-aged children

John C. Thorne, Ph.C., CCC-SLP & Truman Coggins, PhD, CCC-SLP:

with prenatal alcohol exposure

University of Washington Child Language Laboratory

Damage in FASD and

Fetala Alcohol

Spectrum

TREIN and Evidence

during

narrative

generation CNS

task.

Caudate of Basal Ganglia & Frontal Lobes form a functional loop

Part of the Executive system

Match response to context

Predict reward for response

Should impact higher-level aspects of discourse production

More severe impairment in children with FAS facial phenotype expected

Discourse Behaviors

Disorders in Children

(FASD)

of CNS Damage?

Primary Questions:

1)Do TREIN errors increase with increasing evidence of structural CNS

abnormality?

(e.g., Alexander, 2006; Casey, Tottenham, & Fossella, 2002; Cools, Ivry, & D'Esposito, 2006; Longworth, Keenan, Barker, Marslen-Wilson, & Tyler, 2005; Monchi, Petrides, Strafella, Worsley, & Doyon, 2006; Poldrack

& Rodriguez, 2004; Robles, Gatignol, Capelle, Mitchell, & Duffau, 2005; Seger, 2006; Seger & Cincotta, 2006; Teichmann, Dupoux, Kouider, & Bachoud-Levi, 2006; Teichmann et al., 2005; Tricomi, Delgado,

McCandliss, McClelland, & Fiez, 2006; Bitan et al., 2006; Blakemore, 2008; Conklin, Luciana, Hooper, & Yarger, 2007; Ofen et al., 2007; Santesso & Segalowitz, 2008).

Analysis of Narrative Generation during Frog Where Are You?

Tallying Reference Errors In Narrative:TREIN (Thorne 2006; http://students.washington.edu/jct6)

Secondary Question: What is the relationship between severity of FAS diagnosis

and narrative performance when additional narrative measures are included

in analysis?

TREIN quantifies 9 aspects of Narrative:

Appropriate reference strategies:

Indefinite Nominal Introduction

Definite Nominal Introduction

Possessive Nominal Introduction

Static Encephalopathy = CNS RANK (3) or (4)

60% >2 SD smaller 1 or more regions

Smaller caudate nucleus of basal ganglia

Absolute and relative to brain size

about 40% caudate nucleus > 2 SD smaller

Neurobehavioral Disorder = CNS RANK (2)

43% >2 SD smaller 1 or more regions

CNS RANKS of 32 School-age Children:

16 diagnosed with FASD:

CNS RANKS

1= unlikely

none

2 = possible

(n=8)

3 = probable

(n=4)

4 = definite

(n=4)

16 typically developing (TD) age-matched

All considered to have CNS-rank 1

Cohesive Nominal Reference TIE

Cohesive Pronominal Reference TIE

Risk of CNS damage increases with CNS Rank in FASD

FAS facial phenotype

70% >2 SD smaller 1 or more

regions

-Smaller frontal lobes as group

-Absolute and relative

-33% >2 SD smaller

(Astley et al. Forthcoming)

Nominal Reference Errors:

Ambiguous Nominal Introduction

Ambiguous Nominal Reference TIE

from Frog Where Are You? copyright Mayer 1969

Rate of Nominal Reference Errors

.07

CNS RANK

INI

rNRE

mean (SD)

16.7 (3.6)

Significant Contrasts

1-unlikely (n=16)

1>4 (p=0.003)

mean (SD)

1.57% (0.7)

2-possible (n=8)

15.4 (3.5)

2>4 (p=0.03)

2.59% (0.9)

2<4 (p=0.01)
3-probable (n=4)
14.0 (2.4)
ns
3.40% (1.2)
3>1 (p=0.006)

4-definite (n=4)

8.8 (2.8)

4<2(p=0.03);
4<1(p=0.003)
4.48% (0.9)
Significant Contrasts
1<4 (p<0.0001);
4>1 (p<0.0001);
1<3 (p=0.006)
4>2 (p=0.01)

Primary Question 2: Sensitivity/Specificity

Only rNRE had clinically important ability to identify Static Encephalopathy (Plante & Vance, 1994)

2 sd cut-off

sensitivity 88%,

specificity 92%

All False-positives CNS-Rank 2

1.5sd and 1.25 sd

sensitivity 88%

specificity (79% & 71% respectively)

False-positives included TD children.

Secondary Question: Correlations

Significant Correlations to CNS-Rank (Kendalls tau b, p< 0.05)
Measure
Rate of Nominal Reference Error (rNRE)
Indefinite nominal introduction (INI)
Cohesive pronoun ties
Total Utterances
Number of Different Words (NDW)
Number of Words
Personal pronouns
Correlation
0.63
-0.40
-0.37
0.37
-0.31
-0.31
-0.28
Coding conducted blind to storyteller
Interrater reliability: Kappa = 0.90
Conclusions and
Moving Forward:
Appropriate Introductions
30
There was a significant increase in TREIN errors with increasing evidence of
structural CNS damage in this group of 32 children.
.05
rNRE (p=0.0001; ANOVA)
Pairwise Comparisons
(Scheffe; p<0.05)
9 TREIN measures (TREIN Thorne, 2006).
Secondary Question: Correlational analysis (Kendals Tau b)
.06
Groups defined on CNS Rank differed in mean rNRE (F= 15.8, p= 0.0001) and in their mean
number of Indefinite Nominal Introductions (INI; F=5.9, p= 0.003), a near-reciprocal of NRE.
Productivity: NDW, MLU, length measures, etc.
Mazes: Total Maze Words, per Utterance, etc.
Standard Words: Personal Pronouns, etc.
Primary Question 2: Calculate Sensitivity and Specificity for Static
Encephalopathy at 2, 1.5, 1.25 SD from the mean of TD group
(Thorne & Coggins 2008)
rate of Nominal Reference Errors (rNRE = NRE/Total Words)
- Identifies storyteller as FASD vs. Typically Developing Peer (TD)
- Can distinguish storytellers with FAS from other FASD
Primary Question 1: Significant effects of CNS-rank on 2 of 9 TREIN measures.
Narrative measures analyzed
26 standard SALT measures (Miller, 2004).
Analysis:
Primary Question 1: One-way ANOVA with Scheffes contrasts
Pronominal Reference Errors
Ambiguous Pronominal Reference TIE
Pronominal Introduction
Results
2) Do significant differences (i.e., > 2 SD) in narrative performance discriminate

between FASD with Static Encephalopathy and TD peers?

20

These errors were best reflected in the rate of Nominal Reference Errors

(rNRE), demonstrating clinical potential for the measure.

.04

.03

Indefinite Nominal

10

.02

(p= 0.003; ANOVA)

.01

Definite Nominal

0.00

0

unlikely

possible

probable

Increasing Evidence of CNS damage

definite

Replication in a larger, more representative sample of children would bolster

validity of using the TREIN to identify integrative language impairment

resulting from CNS damage in this population.

Possessive

unlikely

possible

probable

definite

Increasing Evidence of CNS damage

Additional findings:

- Children with FAS had highest rNRE as expected (4.36%, SD 0.8)

- children with TD had lowest rNRE (1.57%, SD 0.7)

We are currently working to replicate these results using 160 Frog Where Are

You narratives available from the University of Washington Fetal Alcohol

Syndrome Diagnosis and Prevention Network.

Future research will directly examine structural CNS abnormalities in

children with FASD that exhibit difficulties with narrative generation as

measured by the TREIN.

- children with FASD without FAS fell in between (2.77%, SD 0.9)

(F= 23.0, p< 0.0001; contrasts: FAS>TD, p<0.0001, FAS>other FASD, p=0.005, other FASD>TD, p=0.004)

All 5 children with FAS and 4 with other FASD had rNRE > 2 SD from the TD group mean.

Acknowledgements:

Dr. Lesley Olswang, Brianne Bowker and everyone at the Child Language Laboratory

p-value

<0.0001
=0.006
=0.008
=0.010
=0.025
=0.028
=0.043
Those with Static Encephalopathy had significantly smaller NDW (mean 92.1, SD 25.8)
than the group without (mean 125.8, SD 38.6; t-test -2.29, p=0.03).
Those with Static Encephalopathy also produced fewer cohesive pronoun ties (PNTIE=16, SD 8.6)
than the group without (mean 24.4, SD 11.6; t-test -2.18, p=0.04)
No cases produced NDW or PNTIE > 2 SD below mean of TD group (NDW=130, SD 43.1; PNTIE=27.4, SD 12).

No other measures analyzed produced important contrast between groups.

The whole crew at the UW FASDPNData provided by "Intervening with Children/Adolescents with FAS/ARND"

Susan J. Astley, Ph.D., Principal Investigator; Heather Carmichael Olson, Ph.D., CoInvestigator: Agency:

Centers for Disease Control and Prevention - U.S. Department of Health and Human Services

Funding:

UWSPHSC Research Training in Speech, Language, and Hearing Sciences; Lynn Werner, Principal

Investigator, NIH

LEND: "Interdisciplinary Leadership Training in Neurodevelopmental and Related

McLaughlin, M.D., Principal Investigator, HRSA MCH Bureau

Disabilities" John F.