Stuttering treatment within an operant framework

1
Stuttering treatment within an operant framework

• HESP 612

2
Background

• In most behavioral treatments, positive
  reinforcement (reward) is the most commonly
  identified operant contingency.
• However, stuttering has most often been examined
  in relationship to punishment.
• The most common punishment principle that has
  been used is time-out from reinforcement (TO),
  also called response-contingent time-out (RCTO).

3
Contingent responses and behavior
4
Efficacy of TO

• Most classic laboratory studies are somewhat old
• Haroldson, Martin Starr (1968) 4 adults,
  stuttering was reduced by 88.
• Early replications and extensions
• Martin Berndt (1970) 1 boy
• Martin, Kuhl Haroldson (1972) (the Puppet
  study) 2 preschoolers talked with a puppet and
  received RCTO, stuttering dropped to near zero
  and gains were maintained at one-year follow-up.
• Martin Haroldson (1979) for 20 adult PWS, TO
  was more effective than RC noise, wrong (neg
  verbal feedback, RC DAF, or paced speech. Average
  reduction in stuttering was 76.

5
Early variations on the theme

• Can PWS in group environments administer the tx?
• Martin Haroldson (1971) suggested the system
  did not work well when PWS monitored and
  reinforced each other.
• Egolf, Shames Selzer (1971) rotating speaking
  turns in a group as a form of RCTO can reduce
  stuttering.
• Improving generalization self-administration of
  TO may lead to greater gains than
  clinician-administered TO (Martin Haroldson,
  1972).

6
Fine-tuning TO

• Length of the TO does NOT seem to affect its
  efficacy, although 30 sec. appears marginally
  most efficacious (James Ingham, 1974)
• Contingency of TO is necessary non-contingent TO
  doesnt work, suggesting the TO is not merely a
  distractor (James, 1981).

7
Effects of other forms of punishment on fluency

• Shock variable effects, although group trends
  suggest reduction of stuttering (see Costello,
  1984)
• DAF a classic reducer of stutter events
  (Goldiamond, 1962,1965)
• Verbal contingencies all work, suggesting that
  they highlight moments of stuttering and punish
  regardless of content. The classic examples are
  Cooper, Cady Robbins (1970) and Daly Kimbarow
  (1978) CC R found that right, wrong and tree
  ALL reduced stuttering in adults and children.

8
Can stuttering be operantly INCREASED?

• The test of a behavior truly responsive to
  operant techniques is to see if it can be
  increased through positive reinforcement.
• Virtually no studies of adults have managed to
  increase stutter events via token or monetary
  award systems.
• An unpublished paper by Costello Felsenfeld
  (1979) suggested that the stuttering of a
  six-year old child could be manipulated through
  reward consisting of praise and tokens.

9
Examples of current operant programs for
stuttering ELU

• Costello (Ingham) (1984) and elsewhere Extended
  Length of Utterance (ELU) program
• Contingencies for fluency verbal praise and
  tokens used in token economy
• Steps and criteria samples
• Single fluent word 12 trials 92
• 2 consecutive words 10 100
• 3-6 consecutive words 10 each 100
• 5 sec 90 sec fluent monologue variable
• 2 min 5 min fluent monologue variable

10
The Monterey Program (Ryan and van Kirk Ryan,
1995 and elsewhere)

• Combines DAF and GILCU
• DAF shapes speech from 40 wpm in seven step
  series in reading, monologue, conversation
• GILCU shapes from single word utterances in 54
  steps to five minutes of fluent reading,
  monologue and conversation.
• Both programs highly operant, with verbal and
  token reinforcement for fluency and negative
  consequation for stuttering (e.g., stop, speak
  fluently/in your fluent pattern).

11
The Lidcombe Programme

• Onslow et al. - parent-administered program for
  treatment of early childhood stuttering
• Original contingencies praise on fixed schedule
  for fluency, request for retrial on fixed
  schedule for stutters.
• Revised contingencies praise only.
• Fully documented program with impressive efficacy
  data
• See manual and forms I\COURSES\LidcombeProgramMa
  nual.pdf
• I\COURSES\lidspeech_measures_form.pdf

12
How do operant programs work?

• Some disagreement about establishing the negative
  and positive values of contingencies
• Some contingencies viewed as negative by
  clinicians are viewed positively (and
  conceptually) by clients
• Some evidence that RCTO/punishment is most
  effective when client has prior therapy
  techniques to utilize following TO (James, et
  al., 1974)

13
New data on Lidcombe (Bonelli, et al., 2000)

• Some evidence that fluency gains may be indirect
  result of reinforcement/punishment schedule that
  shapes shorter responses by children.
• Read the articleI\COURSES\Bonellietal.pdf
• Read more about the Lidcombe programme at
  http//www.cchs.usyd.edu.au/asrc//treatment/lidcom
  be.htm

14
How to get children to simplify language without
simplifying parental models one possible notion

• Bonelli, Dixon, Onslow Ratner (CLP, 2000)
  analyzed the operant-based Lidcombe program for
  behaviors that distinguished pre- and
  post-therapy language use by parents and
  children.
• This program has parents praise fluent utterances
  (which are statistically likely to be shorter and
  less ambitious than stuttered utterances)
• The primary finding was that study children began
  treatment with relatively high expressive
  language scores that plateaued (did not meet
  growth expectations) post-therapy.
• The children seemed to intuit that parents were
  happy with shorter, simpler speech turns these
  turns were less likely to stress the developing
  fluency system.

15
Intake/outcome language measures for Lidcombe
children
16
Summary operant procedures

• Strengths of operant approaches
• Strongly data-based, with individual baseline,
  treatment and follow-up data some of the best
  data out there.
• Good short-term and now, evolving long-term
  efficacy data, particularly for very young
  children
• Can be parent-administered, under guidance
  (Lidcombe)
• Limitations of operant approaches
• Not widely used for large-scale clinical
  evaluation programs are confined to small
  numbers of historical working groups
• Follow-up data sometimes lacking
• Assessment and outcomes of the As and Cs not
  usually available.
• Somewhat of a theoretical black box in terms of
  links to stuttering etiology and functional
  source of efficacy (i.e., WHY/HOW DOES IT WORK?)
  AND IF STUTTERING ISNT LEARNED, WHY/HOW should
  it respond to behavioral therapy?