Bound Morpheme Intervention for Children who use AAC Cathy

1
Bound Morpheme Intervention for Children who use
AAC

• Cathy Binger, Ph.D., CCC-SLP
• University of New Mexico
• cbinger_at_unm.edu
• Division 12 Conference
• February 22-24, 2008

2
Overview

• Introduction of the Issues
• Global measures of receptive grammar
• Receptive morphology
• Global measures of expressive grammar
• Expressive morphology
• Bound morpheme intervention study
• Method
• Results
• Discussion

3
Introduction of the Issues Systematic
literature review of grammar SSPI (Binger
Light, in press)

• Global Receptive Grammar Measures (both
  morphology syntax)
• Many children within normal limits for their age
• Many below normal limits for their age
• No consistent differences based on profiles of
  the individuals who used AAC

4

• Receptive Morphology
• Large-scale study found that children with CCN
  had greater difficulty with bound morpheme
  comprehension than receptive-vocabulary matched
  controls (Blockberger Johnston, 2003)
• Other children have demonstrated competencies
  with detecting bound morpheme errors (Redmond
  Johnston, 2001 Sutton Gallagher, 1995)

5

• Global Expressive Grammar Measures
• Tendency to produce brief utterances
• Children with SSPI (e.g., Smith Grove, 1999
  Sutton Gallagher, 1995)
• Children without SSPI who use AAC for research
  purposes (Smith, 1996 Sutton Morford, 1998)
• Many children adolescents with SSPI can and do
  produce longer utterances (e.g., Kelford Smith et
  al., 1989 Lund, 2001)

6

• Expressive Morphology
• Many children who use AAC can and do use a wide
  range of bound morphemes
• Many demonstrate morphological errors
• Individual difference apparent across studies
  (e.g., Blockberger Johnston, 2003 Bruno
  Trembath, 2006 Kelford Smith et al., 1989
  Sutton Gallagher, 1993)

7

• Bound Morpheme Intervention
• Bruno Trembath (2006) targeted subject is
  verb ing
• Some children demonstrated progress others did
  not
• No experimentally controlled studies located
• Need intervention research!
• Overall findings
• Children who use AAC tend to have problems with
  morphology
• Need experimentally controlled intervention
  studies

8
Current Investigation

• Impact of modeling and recasting on bound
  morpheme acquisition
• Adherence with principles of grammar facilitation
  (Fey, Long, Finestack 2003)
• Achieve greater facility with grammar use
• Ensure functional readiness of targeted forms
• Create opportunities to use targets
• Provide systematic recasts
• Avoid telegraphic models

9
Method

• Design
• Single subject multiple baseline across targets
• Heterogeneous population
• (Relatively) low incidence population
• Allows for flexibility with intervention
• Replicated across 2 additional participants

10
Participants

• General criteria
• 3-12 years of age
• Less than 50 comprehensible speech on Index of
  Augmented Speech Comprehensibility in Children
  (Dowden, 1997)
• SGD
• Native English speaker
• Minimum age equivalent of 30 on all TACL-3
  subtests (Carrow-Woolfolk, 1999)
• Fail to correctly use at least 3 different bound
  morphemes
• Hearing vision within or corrected to be within
  functional limits

11

• Assumptions
• Should be developmentally ready to learn
  generalizable grammar rules
• (beyond the item-by-item phase)
• May have a (fairly) complete underlying language
  system
• Indicated by grammar comprehension testing
• (Sidenote use of language comprehension to set
  language expression expectations)

12
Alex

• Age 11
• Ethnicity Hispanic
• Disability CP
• TACL 63 70 age equivalent
• 63 Grammatical Morphemes
• Expressive language
• Short, grammatically incomplete sentences
• Literacy spells some words independently
  invented spellings some word recognition skills
• SGD Vantage

13
Jessie

• Age 6 years
• Ethnicity Anglo/Hispanic
• Disability Childhood apraxia of speech
• TACL 43-49 age equivalent
• 49 Grammatical Morphemes
• Expressive language
• Short, grammatically incomplete sentences
• Literacy id letters sounds cannot spell
  independently
• SGD Vantage

14
Ian

• Age 9
• Race/ethnicity Hispanic African American
• Disability CP
• TACL 50-66 age equivalent
• 50 Grammatical Morphemes (1st percentile)
• Expressive language
• Short, grammatically incomplete sentences
• Literacy id letters and sounds does not spell
  words independently
• SGD Vanguard

15
Materials

• Storybook series chosen by each participant
• Alex Scooby Doo
• Jessie Muppet Babies
• Ian Teenage Mutant Ninja Turtles Spiderman
• SGD
• Whatever device each child already used
• All had Vantage or Vanguard (PRC)
• Vocabulary
• Used vocabulary pre-programmed on
  Vantage/Vanguard

16
Bound Morpheme Representation

• Each bound morpheme represented with MinspeakTM
  icons as they appear on any MinspeakTM-based
  device
• Each morpheme represented with
• Orthographic representation
• Graphic representation

17
Targets

• Alex
• Aux verb main verb -ing (I am running)
• Possessive s (Grandmothers couch)
• 3rd person singular s (he runs)
• Jessie
• Possessive s
• Regular past tense ed (She walked)
• Plural s (shoes)
• Ian
• Regular past tense -ed
• Possessive s
• Aux verb main verb -ing

18
Dependant MeasureProbes

• Percent correct on 10 probes
• Pool of 50 probes for each morpheme
• 10 probes randomly selected and administered at
  the beginning of each session
• Clipart figures used for each probe item
• Figures placed on felt board

19
Probes Aux verb Main verb -ing (McShane
Whitaker, 1988 Leonard et al., 2004)

• Scaredy cat Sometimes I get really scared
  and I hide.
• Oh no, I think theres a rabbit coming and I
  think hes gonna turn
• Researcher brings out rabbit and makes him turn
• Scaredy cat Please tell me whats happening
• Researcher The rabbit says, I...
• Child I AM TURNING

20
Possessive s

• Researcher Here is grandma and here is her
  hamburger. This hamburger belongs to
  grandma. The hamburger is
• Child GRANDMA S

21
3rd Person Singular s (Schütze Wexler, 2000
Leonard et al., 2004)

• Character Here is my pool. Every time Im
  in my pool I swim.
• Researcher What does grandpa do every time hes
  in his pool? He
• Child SWIM S.

22
Plural s

• Researcher Here is one horse here are two
• Child HORSE S

23
Regular Past Tense ed

• Researcher
• The monkey is going to pick his nose.
• Acts out action
• Hes all done picking. What did he do? He
• Child HE PICK ED

24
ProceduresBaseline

• 10 probes administered
• Minimum of 3 baseline sessions
• Stable baseline required
• No intervention provided

25
Procedures Intervention

• 15 minute story reading sessions
• Primary intervention techniques
• Modeling
• Recasting (e.g., Fey, Long, Finestack, 2003)
• At least 10 models and/or recasts per session
  Models and recasts formulated based on events in
  the story

26
ProceduresIntervention

• Models
• Aided AAC models of targets were provided (Binger
  Light, 2007)
• Spoken model
• Model on childs SGD
• E.g., Scooby-Doo is eating Scooby-snacks. HE IS
  EATING.
• Spoken model grammatically complete message
• Aided AAC targeted form

27
Procedures Intervention

• Recasts
• Revised errors
• Alex HE ARE EAT.
• Researcher Lets try, He is eating HE IS
  EATING
• Completed sentences
• Ian THEY
• Researcher THEY KICK ED
• 10 probes administered at the beginning of each
  session to track progress

28
ProceduresMaintenance

• Maintenance probes taken 2, 4, and 8 weeks
  after intervention
• Additional probes taken in some cases
• Alex 4 and 5 months post-intervention

29
ProceduresData Analysis

• Percent correct calculated for all baseline,
  intervention, maintenance probes
• All probe data graphed and visually inspected for
  changed in slope, level, trend (McReynolds
  Kearns, 1983)
• Intervention continued until at least 80
  accuracy for 3 consecutive sessions

30
Results

• All 3 children acquired all 3 bound morphemes
• All 3 children failed to maintain use of the
  first morpheme that was taught
• Jessie fell below criterion for 1st two morphemes

31
Maintenance Errors

• All of the children demonstrated
  overgeneralization
• Alex
• Overgeneralized Aux V on possessive probes
• MAN IS TELEPHONE instead of MAN S TELEPHONE
• Jessie
• Overgeneralized possessive on past tense probes
• MOTHER S KISS instead of MOTHER KISS ED
• Ian
• Overgeneralized possessive on past tense probes
• SHE DRY S instead of SHE DRY ED

32

• Overgeneralization errors
• Crossed word classes
• SHE KISSS (possessive s on a verb)
• Inappropriate context
• GIRLS CLIMB (possessive s before a verb)
• Added 2 morphemes
• HE LAUGH ED S (plural)
• ? All made unusual errors that would not be seen
  with typically developing children

33

• All of the children failed to used previously
  learned morphemes
• Alex
• Did not use Aux V for Aux V Main V ing
• I BUILD ING, THEY STEAL ING
• Jessie
• Did not use possessive s on possessive probes
• CAT SPAGHETTI instead of CAT S SPAGHETTI
• Ian
• Did not use ed on past tense probes
• SHE TALK TO A DOG instead of SHE TALK ED TO A
  DOG

34
Intervention Phase II

• Provided additional intervention for targets that
  fell below criterion
• Used contrastive training (Connell, 1982)
• Contrasted Target 1 with Target 2 for each
  child
• Jessie Also contrasted Target 2 with Target 3

35
Contrastive Intervention

• Contrastive training was successful for all
  children
• All maintained use of bound morphemes after
  contrastive training

36
Maintenance Take II

• All 3 children maintained use of targets after
  contrastive intervention provided

37
Discussion

• All 3 participants (eventually) learned to use
  all 3 targeted morphemes
• Single subject designed allowed for revisions to
  be made with the intervention
• Important for trying out new interventions

38
Effects of Modeling Recasting

• Modeling and recasting appeared to have been an
  effective intervention at first
• All reached criterion
• But all children failed to maintain use of 1st
  morpheme
• Children confused 1st and 2nd bound morphemes
• Jessie also confused 2nd and 3rd

39

• Modeling and recasting may have had a priming
  effect
• Provided initial exposure to the morphemes
• May have been an important precursor to
  contrastive intervention
• Was sufficient for Target 2 (for Alex Ian) and
  Target 3 (for all)

40

• Special note on recasting
• Seemed to be much more salient and corrective
  than when using speech
• Speech
• Auditory feedback
• Aided AAC
• Auditory, visual, tactile feedback

41
What Happened During Maintenance?

• Possibility 1 Did not truly learn the meaning
  of the morphemes at first
• Appeared to learn rote patterns
• E.g., When Cathy asks me to do something, I
  should put in a word ending (s/s/-ed/-ing)
  somewhere
• Could succeed with 1st set of probes in this
  manner
• Could not succeed once additional morphemes
  introduced

42

• Possibility 2 Negative interaction when trying
  to express other new linguistic forms
• Ian seemed to have problems when expanding his
  messages
• SHE TALK TO A DOG
• Frequently failed when expanding utterances
• Seemed anxious to complete the sentence
• Eventually began to self-correct in these
  circumstances

43

• Possibility 3 Intervention was not long enough
• Modeling recasting may have been sufficient had
  intervention extended longer
• Assumes that probes were not effective measure of
  acquisition

44
Potential Issues with Probes

• Probes
• Widely used to evaluate performance on
  grammatical morphemes (e.g., Leonard et al.,
  2004 Schütze Wexler, 2000 McShane Whitaker,
  1988)
• May require revisions
• Participants met criterion on probes initially,
  but they did not seem to really learn the
  morphemes
• May be more useful when different targets are
  probed within same set (not 10 in a row of same
  target)
• May need to include foils (Here is 1 horse and
  here is 1 ___

45
Directions for Future Research

• Many different techniques to evaluate
• Modeling
• Recasting
• Imitation (speech aided AAC)
• Contrastive intervention
• In conjunction w/ modeling recasting
• Forced alternative questions (Is it the BOY WALK
  S or the BOY WALK ED?)
• Correction of incorrect forms
• Similar to recasting when using aided AAC
• Explanation of grammatical rules

46

• Other contexts
• Need to look at generalization to academic and
  possibly certain social contexts
• Balance with efficiency issues
• Other communication partners
• Adapt partner instruction models (Kent-Walsh
  McNaughton, 2005 Binger et al., in press)
• Teach aided AAC modeling, recasting, etc.

47

• Many remaining questions regarding
• Who receives intervention?
• What types of intervention might be appropriate?
• What intervention techniques might be
  appropriate?
• When is intervention appropriate for a person?
• How should grammatical morphemes be represented?
• Well tackle these questions in the Discussion
  Section