THE	
  EFFECTS	
  OF	
  DISTAL	
  PROSODIC	
  CUES	
  ON	
  SPEECH	
  PERCEPTION	
  IN	
  ADULTS	
  
WHO	
  STUTTER	
  
	
  
By	
  
	
  
Kaitlyn	
  Marie	
  Ayres	
  
	
  

	
  
	
  
A	
  THESIS	
  

	
  

Submitted	
  to	
  
Michigan	
  State	
  University	
  
in	
  partial	
  fulfillment	
  of	
  the	
  requirements	
  
for	
  the	
  degree	
  of	
  
	
  
Communicative	
  Sciences	
  and	
  Disorders	
  -­‐	
  Master	
  of	
  Arts	
  
	
  
2015	
  
	
  

ABSTRACT	
  
THE	
  EFFECTS	
  OF	
  DISTAL	
  PROSODIC	
  CUES	
  ON	
  SPEECH	
  PERCEPTION	
  IN	
  ADULTS	
  
WHO	
  STUTTER	
  
	
  
By	
  
Kaitlyn	
  Marie	
  Ayres	
  
Stuttering	
  is	
  a	
  neurodevelopmental	
  disorder	
  characterized	
  by	
  the	
  disruption	
  of	
  
rhythmic	
  flow	
  and	
  timing	
  in	
  speech	
  production.	
  Speech	
  productions	
  of	
  people	
  who	
  
stutter	
  are	
  influenced	
  by	
  the	
  perception	
  of	
  prosodic	
  aspects	
  of	
  their	
  own	
  speech,	
  
suggesting	
  that	
  people	
  who	
  stutter	
  may	
  have	
  difficulty	
  in	
  perceiving	
  prosodic	
  
elements	
  of	
  their	
  own	
  speech	
  causing	
  disfluency.	
  To	
  investigate	
  this	
  possibility,	
  
adults	
  who	
  stutter	
  (n=15)	
  and	
  adult	
  fluent	
  speakers	
  (n=13)	
  were	
  compared	
  in	
  their	
  
ability	
  to	
  perceive	
  distal	
  rhythmic	
  and	
  timing	
  cues	
  in	
  speech.	
  Adults	
  who	
  stutter	
  
exhibited	
  perception	
  of	
  the	
  distal	
  prosodic	
  cues	
  and	
  an	
  ability	
  to	
  internally	
  generate	
  
rhythm	
  and	
  timing	
  for	
  parsing	
  of	
  syllables	
  and	
  words	
  in	
  speech.	
  These	
  findings	
  
provide	
  evidence	
  of	
  intact	
  perception	
  of	
  rhythm	
  and	
  timing	
  of	
  speech	
  in	
  adults	
  who	
  
stutter.	
  	
  
	
  

ACKNOWLEGMENTS	
  
	
  
I	
  would	
  like	
  to	
  thank	
  my	
  thesis	
  committee	
  members,	
  Dr.	
  Laura	
  Dilley,	
  Dr.	
  Soo-­‐Eun	
  
Chang,	
  and	
  Dr.	
  Amanda	
  Hampton	
  Wray	
  for	
  all	
  of	
  their	
  encouragement,	
  support,	
  and	
  
guidance	
  with	
  this	
  research.	
  Additionally,	
  I	
  thank	
  Elizabeth	
  Wieland,	
  Gregory	
  Spray,	
  
Kristin	
  Hicks,	
  Saralyn	
  Rubsam,	
  Zachary	
  Ireland,	
  and	
  other	
  members	
  of	
  the	
  MSU	
  
Speech	
  Perception-­‐Production	
  Laboratory	
  and	
  MSU	
  Speech-­‐Neurophysiology	
  
Laboratory	
  for	
  their	
  generous	
  assistance	
  with	
  this	
  project.	
  	
  
	
  

	
  

	
  

iii	
  

TABLE	
  OF	
  CONTENTS	
  
	
  
LIST	
  OF	
  TABLES…………………………………………………………………..……………………………....v	
  
LIST	
  OF	
  FIGURES…………………………………………………………………………………………………vi	
  
KEY	
  TO	
  ABBREVIATIONS……………………………………………………………………………………vii	
  
Introduction…….…………………………………………………………………………………………………..1	
  
Participants……………………………………………………………………………………………………….11	
  
Experiment	
  1:	
  Distal	
  Intonation	
  Rhythm……………………………………….…………………….13	
  
	
  
Methods……………………………………………………………………….…………………………13	
  
	
  
Results……………………………………………………………………………………………………17	
  
	
  
Discussion………………………………………………………………………………….…..............18	
  
Experiment	
  2:	
  Distal	
  Speech	
  Rate……………………………………………….……………………….20	
  
	
  
Methods……………………………………………………………………………….………………....20	
  
	
  
Results……………………………………………………………………………………………………22	
  
	
  
Discussion……………………………………………………………………………………………....24	
  
General	
  Discussion	
  and	
  Conclusions…..……………………………...………………………………..25	
  
APPENDICES………………………………………………………………………………………………….…..33	
  
	
  
Appendix	
  A:	
  Experiment	
  1	
  Items	
  List………………………………………..……………...34	
  
	
  
Appendix	
  B:	
  Experiment	
  2	
  Items	
  List…………………………………………..…………...35	
  
	
  
	
  
REFERENCES………………………………………………………………………………………………...…...36	
  

	
  

iv	
  

LIST	
  OF	
  TABLES	
  
	
  
Table	
  1:	
  Participant	
  Information…………………………………………………………………………12	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

v	
  

LIST	
  OF	
  FIGURES	
  
	
  
Figure	
  1:	
  Distal	
  Context	
  Manipulations……………………………………………..…………………16	
  
Figure	
  2:	
  Proportion	
  of	
  Disyllabic	
  Responses…………………………………………..…………..18	
  
Figure	
  3:	
  Distal	
  Rate	
  Manipulations………………………………………………………...……..……20	
  
Figure	
  4:	
  Reported	
  Function	
  Word	
  by	
  Rate………………………………………………………….23	
  	
  
	
  
	
  

	
  

vi	
  

KEY	
  TO	
  ABBREVIATIONS	
  
	
  
Children	
  who	
  stutter	
  ………………………………………………………………………………………CWS	
  
Persons	
  who	
  stutter	
  …………………………………………………………………………………...…..PWS	
  
	
  
Delayed	
  auditory	
  feedback……………………………………………………………………………….DAF	
  
	
  
Basal	
  ganglia	
  thalamocortical	
  network…………………………………………………………...BGTC	
  	
  
	
  

	
  

vii	
  

Introduction	
  
	
  

Stuttering	
  is	
  a	
  speech	
  disorder	
  marked	
  by	
  frequent	
  hesitations,	
  

prolongations	
  and/or	
  repetitions	
  of	
  sounds,	
  syllables	
  and	
  words.	
  	
  These	
  types	
  of	
  
stuttering	
  disfluencies	
  characterize	
  the	
  disorder	
  as	
  a	
  disruption	
  of	
  the	
  rhythmic	
  flow	
  
of	
  speech	
  production	
  (World	
  Health	
  Organization,	
  2010).	
  Although	
  many	
  theoretical	
  
models	
  suggest	
  possible	
  underlying	
  mechanisms	
  or	
  causes	
  of	
  stuttering,	
  none	
  have	
  
thus	
  proven	
  to	
  provide	
  a	
  completely	
  sufficient	
  explanation	
  (e.g.,	
  Packman,	
  Code,	
  &	
  
Onslow,	
  2007).	
  	
  
	
  

All	
  components	
  of	
  language	
  production,	
  including	
  pitch	
  and	
  rhythm	
  

characteristics	
  (referred	
  to	
  as	
  prosody),	
  have	
  been	
  a	
  recurring	
  focus	
  in	
  this	
  search	
  
for	
  a	
  cause	
  of	
  stuttering.	
  	
  Various	
  previous	
  research	
  has	
  shown	
  that	
  aspects	
  of	
  
prosody	
  play	
  important	
  roles	
  in	
  the	
  production	
  and	
  perception	
  of	
  fluent	
  speech	
  
(Bergmann,	
  1986;	
  Besozzi	
  &	
  Adams,	
  1969;	
  Hall,	
  Amir,	
  &	
  Yairi,	
  1999;	
  Packman	
  et	
  al.,	
  
2007;	
  Packman,	
  Onslow,	
  Richard,	
  &	
  Van	
  Doorn,	
  1996;	
  Wingate,	
  1966).	
  The	
  known	
  
symptoms	
  of	
  dysrhythmic	
  flow	
  in	
  the	
  motor	
  speech	
  production	
  of	
  persons	
  who	
  
stutter	
  (PWS)	
  have	
  led	
  to	
  proposals	
  that	
  an	
  underlying	
  deficit	
  in	
  PWS	
  may	
  be	
  
associated	
  with	
  the	
  prosodic	
  aspects	
  of	
  rhythm	
  and	
  timing	
  used	
  in	
  speech	
  (Alm,	
  
2004).	
  	
  	
  Stuttering	
  has	
  been	
  compared	
  to	
  other	
  disorders	
  such	
  as	
  dystonia	
  or	
  
Parkinson’s	
  disease,	
  due	
  to	
  the	
  similarity	
  in	
  symptoms	
  of	
  disrupted	
  rhythm	
  or	
  
timing	
  in	
  motor	
  movements.	
  These	
  disorders	
  are	
  classified	
  as	
  basal	
  ganglia	
  motor	
  
disorders,	
  and	
  thus	
  it	
  has	
  been	
  suggested	
  that	
  an	
  impaired	
  ability	
  of	
  the	
  basal	
  
ganglia	
  in	
  generating	
  timing	
  cues	
  for	
  initiating	
  motor	
  speech	
  is	
  a	
  core	
  dysfunction	
  in	
  
stuttering	
  (Alm,	
  2004).	
  

	
  

1	
  

	
  

Despite	
  not	
  having	
  revealed	
  the	
  underlying	
  deficit	
  of	
  stuttering,	
  research	
  has	
  

identified	
  various	
  conditions	
  in	
  which	
  stuttering	
  is	
  decreased,	
  most	
  of	
  which	
  utilize	
  
aspects	
  of	
  rhythm	
  and	
  timing.	
  Studies	
  have	
  shown	
  that	
  speech	
  productions	
  of	
  
people	
  who	
  stutter	
  (PWS)	
  can	
  be	
  influenced	
  by	
  the	
  perception	
  of	
  prosodic	
  aspects	
  of	
  
their	
  own	
  speech	
  (Hargrave,	
  Kalinowski,	
  Stuart,	
  Armson,	
  &	
  Jones,	
  1994;	
  Wingate,	
  
1966).	
  	
  Packman	
  and	
  colleagues	
  (1996)	
  theorized	
  that	
  the	
  beneficial	
  effects	
  of	
  
rhythmic	
  speech	
  on	
  stuttering	
  are	
  due	
  to	
  the	
  reduced	
  variability	
  of	
  syllabic	
  stress	
  
provided	
  by	
  the	
  external	
  rhythm	
  (i.e.,	
  the	
  V	
  Model).	
  	
  The	
  patterns	
  of	
  syllabic	
  stress	
  
are	
  influenced	
  by	
  the	
  rhythmic	
  patterns	
  of	
  speech	
  perception	
  in	
  PWS.	
  	
  Kent	
  (1984)	
  
suggested	
  that	
  the	
  difference	
  in	
  ability	
  to	
  generate	
  rhythmic	
  patterns	
  for	
  speech	
  
production	
  is	
  a	
  principal	
  distinction	
  between	
  people	
  who	
  stutter	
  and	
  fluent	
  
speakers.	
  These	
  findings	
  throughout	
  various	
  research	
  have	
  been	
  supported	
  by	
  
conditions	
  in	
  which	
  fluency	
  is	
  enhanced,	
  including	
  choral	
  reading	
  (Adams	
  &	
  Ramig,	
  
1980;	
  Ingham	
  &	
  Carroll,	
  1977),	
  singing	
  (Glover,	
  Kalinowski,	
  Rastatter,	
  &	
  Stuart,	
  
1996),	
  and	
  speaking	
  in	
  time	
  with	
  a	
  metronome	
  (Wingate	
  &	
  Howell,	
  2002).	
  These	
  
evidence-­‐based,	
  fluency-­‐inducing	
  conditions	
  all	
  share	
  one	
  key	
  aspect:	
  the	
  provision	
  
of	
  an	
  external	
  rhythm	
  or	
  external	
  pacing	
  of	
  speech.	
  The	
  fluency-­‐inducing	
  effects	
  of	
  
an	
  external	
  pacing	
  signal	
  suggest	
  PWS	
  may	
  have	
  an	
  impaired	
  internal	
  pacing	
  signal.	
  
Knowing	
  that	
  elements	
  of	
  speech	
  productions	
  are	
  generated	
  from	
  the	
  stored	
  
perceptions	
  of	
  speech	
  (Guenther,	
  1994),	
  we	
  can	
  infer	
  that	
  if	
  PWS	
  exhibit	
  a	
  deficit	
  in	
  
perceiving	
  rhythmic	
  prosodic	
  cues	
  in	
  speech,	
  it	
  may	
  be	
  affecting	
  their	
  ability	
  to	
  
internally	
  generate	
  these	
  prosodic	
  cues	
  in	
  their	
  speech	
  productions.	
  	
  

	
  

2	
  

	
  

	
  In	
  this	
  study	
  we	
  investigated	
  the	
  possibility	
  of	
  PWS	
  exhibiting	
  an	
  impaired	
  

utilization	
  of	
  prosodic	
  elements	
  in	
  the	
  perception	
  of	
  their	
  own	
  speech,	
  relating	
  to	
  
their	
  disfluency	
  in	
  speech	
  production.	
  Any	
  deficits	
  found	
  in	
  the	
  perceptions	
  of	
  
prosodic	
  characteristics	
  of	
  speech	
  (pitch	
  rhythm	
  and	
  timing)	
  by	
  PWS	
  support	
  the	
  
idea	
  that	
  disruption	
  of	
  internal	
  rhythm	
  generation	
  from	
  perception	
  of	
  rhythm	
  and	
  
timing	
  in	
  their	
  own	
  speech	
  could	
  be	
  a	
  contributing	
  factor	
  to	
  the	
  core	
  deficits	
  of	
  
stuttering.	
  The	
  existence	
  of	
  deficits	
  in	
  the	
  rhythmic-­‐prosodic	
  components	
  of	
  speech	
  
for	
  PWS	
  are	
  supported	
  by	
  literature	
  which	
  shows	
  treatments	
  involving	
  external	
  
pacing	
  help	
  improve	
  their	
  fluency	
  of	
  speech	
  (Ingham	
  &	
  Carroll,	
  1977;	
  Wingate	
  &	
  
Howell,	
  2002;	
  Glover,	
  Kalinowski,	
  Rastatter,	
  &	
  Stuart,	
  1996).	
  If	
  PWS	
  exhibit	
  a	
  deficit	
  
in	
  perceiving	
  these	
  rhythmic	
  prosodic	
  cues	
  in	
  speech,	
  it	
  may	
  be	
  affecting	
  the	
  ability	
  
to	
  internally	
  generate	
  rhythmic	
  pacing	
  of	
  speech,	
  which	
  would	
  be	
  evident	
  in	
  both	
  
the	
  perception	
  and	
  production	
  of	
  speech.	
  	
  
	
  

Recent	
  findings	
  from	
  functional	
  and	
  structural	
  magnetic	
  resonance	
  imaging	
  

studies	
  have	
  shown	
  neurological	
  evidence	
  of	
  these	
  possible	
  deficits	
  in	
  internal	
  beat	
  
generation	
  that	
  are	
  used	
  for	
  the	
  rhythmic	
  production	
  of	
  speech	
  (Chang	
  &	
  Zhu,	
  2013;	
  
Grahn	
  &	
  McAuley,	
  2009;Toyomura	
  et	
  al.,	
  2015).	
  Rhythm	
  processing	
  and	
  internal	
  
generation	
  of	
  a	
  beat	
  are	
  neurologically	
  supported	
  by	
  the	
  basal	
  ganglia	
  
thalamocortical	
  (BGTC)	
  network,	
  which	
  involves	
  connectivity	
  among	
  the	
  basal	
  
ganglia,	
  supplementary	
  motor	
  area	
  (SMA),	
  and	
  pre-­‐motor	
  and	
  auditory	
  regions	
  
(Grahn	
  &	
  McAuley,	
  2009).	
  These	
  cortical	
  and	
  sub-­‐cortical	
  regions	
  are	
  commonly	
  
activated	
  in	
  tasks	
  that	
  involve	
  motor	
  timing	
  (such	
  as	
  speech)	
  and	
  are	
  strongly	
  
associated	
  with	
  temporal	
  perceptions	
  and	
  processing	
  of	
  auditory	
  stimuli	
  

	
  

3	
  

(Harrington	
  et	
  al.,	
  1998b;	
  Schubotz	
  et	
  al.,	
  2000;	
  Rao	
  et	
  al.,	
  2001;	
  Nenadic	
  et	
  al.,	
  2003;	
  
Coull,	
  2004;	
  Grahn	
  and	
  Brett,	
  2007;	
  Hazeltine	
  et	
  al.,	
  1997;	
  Harrington	
  and	
  Haaland,	
  
1999).	
  Grahn	
  &	
  McAuley	
  (2009)	
  support	
  these	
  notions	
  by	
  finding	
  greater	
  activity	
  of	
  
the	
  pre-­‐motor	
  and	
  supplementary	
  motor	
  cortices	
  and	
  basal	
  ganglia	
  in	
  adults	
  while	
  
making	
  timing	
  judgments	
  in	
  an	
  auditory	
  perception	
  task.	
  	
  Additionally,	
  Chang	
  and	
  
Zhu	
  (2013)	
  showed	
  that	
  children	
  who	
  stutter	
  have	
  attenuated	
  connectivity	
  within	
  
the	
  BGTC	
  network	
  compared	
  to	
  age-­‐matched	
  controls.	
  Toyomura	
  et	
  al.	
  (2015)	
  
suggested	
  that	
  the	
  basal	
  ganglia	
  is	
  involved	
  in	
  the	
  improvement	
  of	
  speech	
  fluency	
  by	
  
utilizing	
  the	
  perceptions	
  of	
  an	
  externally	
  triggered	
  rhythm.	
  	
  A	
  deficit	
  in	
  rhythm	
  
perception	
  was	
  found	
  in	
  children	
  who	
  stutter	
  (CWS)	
  by	
  Wieland,	
  McAuley,	
  Dilley,	
  
and	
  Chang	
  (2015),	
  also	
  providing	
  evidence	
  for	
  a	
  deficit	
  in	
  speech	
  perceptions	
  of	
  
PWS.	
  	
  Differences	
  between	
  internally	
  generated	
  pacing	
  for	
  speech	
  and	
  externally-­‐
triggered	
  pacing	
  may	
  be	
  important	
  for	
  understanding	
  the	
  underlying	
  
neuromechanisms	
  of	
  disfluency	
  in	
  persons	
  who	
  stutter	
  based	
  on	
  the	
  fact	
  that	
  
stuttering	
  occurs	
  in	
  self-­‐paced	
  speech	
  but	
  is	
  temporarily	
  decreased	
  with	
  the	
  
presence	
  of	
  an	
  external	
  pacing	
  (Toyomura,	
  2015).	
  
	
  

All	
  of	
  these	
  findings	
  provide	
  increasing	
  evidence	
  for	
  a	
  possible	
  deficit	
  in	
  the	
  

BGTC	
  network	
  in	
  PWS	
  that	
  could	
  potentially	
  be	
  affecting	
  the	
  ability	
  to	
  internally	
  
generate	
  rhythm	
  and	
  timing	
  for	
  motor	
  production.	
  These	
  results	
  also	
  suggest	
  the	
  
correlation	
  between	
  a	
  possible	
  deficit	
  in	
  the	
  internal	
  rhythm	
  generation	
  within	
  the	
  
self-­‐perceptions	
  of	
  speech	
  by	
  people	
  who	
  stutter	
  with	
  concomitant	
  effects	
  on	
  the	
  
fluency	
  of	
  speech	
  production.	
  

	
  

4	
  

	
  

Although	
  previous	
  research	
  has	
  alluded	
  to	
  the	
  possibility	
  of	
  PWS	
  exhibiting	
  

deficits	
  in	
  prosodic	
  elements,	
  none	
  have	
  before	
  looked	
  at	
  these	
  effects	
  on	
  the	
  speech	
  
perception	
  of	
  PWS.	
  	
  Further,	
  how	
  the	
  perception	
  of	
  these	
  prosodic	
  cues	
  affects	
  
linguistic	
  function	
  in	
  segmentation	
  and	
  syllable	
  parsing	
  has	
  never	
  been	
  investigated	
  
in	
  PWS.	
  In	
  the	
  past	
  35	
  years,	
  research	
  focusing	
  on	
  the	
  influence	
  of	
  prosody	
  on	
  
speech	
  perception	
  targeted	
  how	
  prosodic	
  properties	
  of	
  local	
  pitch	
  and	
  timing	
  
(proximal	
  to	
  a	
  word	
  itself)	
  affected	
  the	
  segmentation	
  and	
  rhythmic	
  properties	
  of	
  
that	
  word	
  (Cutler,	
  Dahan,	
  van	
  Donselaar,	
  1997).	
  	
  Prosodic	
  characteristics	
  of	
  spoken	
  
language	
  influencing	
  the	
  processing	
  and	
  parsing	
  of	
  speech	
  have	
  been	
  shown	
  in	
  
control	
  populations	
  across	
  many	
  different	
  languages.	
  	
  The	
  literature	
  has	
  shown	
  that	
  
listeners	
  tend	
  to	
  perceive	
  stressed	
  syllables	
  (i.e.,	
  syllables	
  that	
  include	
  high	
  pitch	
  
and	
  longer	
  duration)	
  as	
  the	
  beginning	
  of	
  new	
  words	
  and	
  likewise,	
  syllable	
  
lengthening	
  as	
  signaling	
  the	
  end	
  of	
  a	
  word	
  or	
  phrase	
  (Cutler,	
  Dahan,	
  &	
  van	
  
Donselaar,	
  1997;	
  Salverda	
  et	
  al.,	
  2007.	
  	
  While	
  research	
  on	
  the	
  effects	
  of	
  prosody	
  on	
  
the	
  parsing	
  of	
  speech	
  began	
  with	
  investigation	
  of	
  proximal	
  (local)	
  prosodic	
  
characteristics,	
  research	
  has	
  evolved	
  into	
  including	
  investigation	
  of	
  distal	
  (non-­‐
local)	
  prosody	
  on	
  the	
  parsing	
  of	
  speech	
  (Cutler,	
  Dahan,	
  van	
  Donselaar,	
  1997;	
  Dilley	
  
&	
  McAuley,	
  2008).	
  Valid	
  and	
  reliable	
  ways	
  to	
  gather	
  measures	
  of	
  how	
  people	
  
perceive	
  particular	
  elements	
  of	
  distal	
  (non-­‐adjacent)	
  prosodic	
  cues	
  has	
  been	
  
established	
  and	
  tested	
  with	
  typical	
  control	
  populations	
  (Dilley	
  &	
  McAuley,	
  2008;	
  
Dilley	
  &	
  Pitt,	
  2010;	
  Morrill,	
  Dilley,	
  and	
  McAuley,	
  2014).	
  	
  Results	
  from	
  these	
  studies	
  
concluded	
  that	
  typical-­‐developing	
  adults	
  utilize	
  distal	
  (i.e.,	
  distant,	
  non-­‐adjacent)	
  
prosodic	
  cues	
  of	
  rhythm	
  and	
  timing	
  of	
  speech	
  in	
  their	
  parsing	
  of	
  syllables	
  and	
  lexical	
  

	
  

5	
  

segmentation	
  of	
  words.	
  	
  This	
  was	
  demonstrated	
  by	
  the	
  fact	
  that	
  alterations	
  in	
  distal	
  
prosodic	
  cues	
  yielded	
  significantly	
  different	
  speech	
  perceptions	
  for	
  listeners	
  of	
  
proximal	
  material	
  (Dilley	
  and	
  McAuley,	
  2008;	
  Dilley	
  and	
  Pitt,	
  2010;	
  Morrill	
  et	
  al.,	
  
2014).	
  	
  	
  
	
  

Dilley	
  and	
  Pitt	
  (2010)	
  demonstrated	
  clear	
  evidence	
  that	
  a	
  manipulation	
  of	
  

distal	
  speech	
  rate	
  induced	
  typically-­‐developing	
  listeners	
  to	
  segment	
  words	
  
differently.	
  The	
  speech	
  perception	
  task	
  utilized	
  in	
  this	
  study	
  involved	
  judgments	
  on	
  
sentences	
  that	
  included	
  a	
  target	
  syllable	
  with	
  a	
  function	
  word	
  (e.g.,	
  or	
  in	
  a	
  sentence	
  
like	
  Deena	
  didn’t	
  have	
  any	
  leisure	
  or	
  time),	
  presented	
  auditorily.	
  When	
  the	
  distal	
  
speech	
  rate	
  was	
  at	
  a	
  normal	
  rate,	
  listeners	
  reported	
  hearing	
  the	
  function	
  word.	
  
However,	
  when	
  the	
  distal	
  speech	
  rate	
  was	
  slowed,	
  listeners	
  did	
  not	
  perceive	
  the	
  
function	
  word.	
  Despite	
  the	
  fact	
  that	
  the	
  acoustic	
  properties	
  of	
  the	
  proximal,	
  targeted	
  
syllable	
  (function	
  word)	
  of	
  each	
  sentence	
  being	
  acoustically	
  identical,	
  the	
  
perceptions	
  of	
  the	
  target	
  syllable	
  were	
  altered	
  based	
  on	
  the	
  manipulated	
  rate	
  of	
  the	
  
distal	
  context.	
  This	
  finding	
  demonstrates	
  that	
  listeners	
  from	
  a	
  typically	
  developing	
  
control	
  population	
  are	
  sensitive	
  to	
  the	
  distal	
  speech	
  rate.	
  These	
  manipulations	
  also	
  
demonstrated	
  the	
  function	
  of	
  prosodic	
  cues	
  in	
  speech	
  on	
  the	
  lexical	
  segmentation	
  of	
  
speech	
  perception	
  in	
  typical-­‐developing	
  adults.	
  
	
  

Other	
  research	
  has	
  revealed	
  additional	
  sensitivities	
  in	
  typically-­‐developing	
  

listeners	
  to	
  aspects	
  of	
  distal	
  prosody.	
  For	
  example,	
  a	
  paradigm	
  developed	
  by	
  Dilley	
  
and	
  McAuley	
  (2008)	
  and	
  extended	
  by	
  Morrill,	
  Dilley,	
  and	
  McAuley	
  (2014)	
  revealed	
  
effects	
  of	
  distal	
  rhythmic	
  cues	
  on	
  linguistic	
  parsing	
  of	
  syllables	
  into	
  word	
  units.	
  In	
  
these	
  studies,	
  participants	
  heard	
  auditory	
  speech	
  sequences	
  containing	
  two	
  

	
  

6	
  

disyllabic,	
  non-­‐compound	
  words	
  continued	
  with	
  a	
  string	
  of	
  four	
  syllables	
  (e.g.	
  
banker	
  helpful	
  [tɑɪ	
  mɝ˞	
  dɝ	
  bi]).	
  The	
  string	
  of	
  final	
  four	
  syllables	
  had	
  ambiguous	
  
lexical	
  structure	
  and	
  therefore	
  could	
  be	
  organized	
  more	
  than	
  one	
  way	
  (e.g.,	
  timer	
  
derby,	
  tie	
  murder	
  bee).	
  The	
  organization	
  of	
  the	
  final	
  four	
  syllables	
  was	
  shown	
  in	
  
these	
  studies	
  to	
  be	
  dependent	
  on	
  the	
  fundamental	
  frequency	
  (F0)	
  and	
  duration	
  of	
  
the	
  distal	
  context.	
  These	
  studies	
  used	
  sentences	
  such	
  as	
  banker	
  helpful	
  [tɑɪ	
  mɝ˞	
  dɝ	
  
bi],	
  in	
  which	
  the	
  lexical	
  organization	
  of	
  the	
  final	
  four	
  syllables	
  is	
  ambiguous	
  (timer	
  
derby	
  or	
  tie	
  murder	
  bee).	
  The	
  experiments	
  utilized	
  two	
  distinct	
  distal	
  prosodic	
  
conditions	
  that	
  were	
  expected	
  to	
  lead	
  to	
  perceptual	
  grouping	
  of	
  syllables	
  in	
  two	
  
different	
  ways:.	
  In	
  one	
  distal	
  prosodic	
  pattern,	
  a	
  repeated	
  low	
  (L)-­‐high	
  (H)	
  
alternation	
  of	
  F0	
  on	
  distal	
  syllables	
  was	
  predicted	
  to	
  lead	
  to	
  grouping	
  of	
  L	
  and	
  H	
  
pitch	
  elements	
  across	
  the	
  entire	
  sequence,	
  which	
  was	
  predicted	
  to	
  generate	
  
perception	
  of	
  a	
  disyllabic	
  word	
  (e.g.,	
  derby);	
  these	
  were	
  referred	
  to	
  as	
  Disyllabic	
  
contexts.	
  In	
  another	
  distal	
  prosodic	
  pattern,	
  a	
  repeated	
  high-­‐low	
  alternation	
  of	
  F0	
  
was	
  predicted	
  to	
  lead	
  to	
  grouping	
  of	
  H	
  and	
  L	
  pitch	
  elements	
  across	
  the	
  entire	
  
sequence,	
  which	
  was	
  predicted	
  to	
  generate	
  perception	
  of	
  a	
  monosyllabic	
  final	
  word	
  
(e.g.,	
  bee);	
  these	
  were	
  referred	
  to	
  as	
  Monosyllabic	
  contexts.	
  In	
  more	
  detail,	
  for	
  both	
  
the	
  monosyllabic	
  and	
  disyllabic	
  contexts,	
  the	
  final	
  three	
  syllables	
  of	
  each	
  target	
  
sequence	
  received	
  a	
  high	
  (H),	
  low	
  (L),	
  high	
  (H)	
  F0	
  pattern	
  (one	
  F0	
  target	
  per	
  
syllable).	
  The	
  first	
  five	
  syllables	
  of	
  each	
  target	
  sequence	
  varied	
  in	
  F0	
  pattern.	
  The	
  
first	
  five	
  syllables	
  of	
  the	
  disyllabic	
  context	
  items	
  received	
  a	
  L1-­‐H2-­‐L3-­‐H4-­‐L5-­‐	
  
pattern,	
  with	
  one	
  F0	
  target,	
  H	
  or	
  L,	
  on	
  each	
  syllable.	
  The	
  monosyllabic	
  pattern	
  

	
  

7	
  

received	
  a	
  H1-­‐L2-­‐H3-­‐L4-­‐	
  HL5-­‐	
  pattern	
  with	
  one	
  F0	
  target	
  for	
  each	
  of	
  the	
  first	
  four	
  
syllables,	
  and	
  a	
  fall	
  in	
  F0	
  from	
  H	
  to	
  L	
  on	
  the	
  fifth	
  syllable.	
  These	
  distal	
  prosodic	
  
context	
  manipulations	
  predicted	
  the	
  perceptual	
  grouping	
  of	
  the	
  final	
  syllables	
  in	
  the	
  
target	
  sequence,	
  in	
  which	
  the	
  Disyllabic	
  context	
  caused	
  a	
  relatively	
  stronger	
  
prosodic	
  boundary	
  to	
  be	
  heard	
  before	
  the	
  penultimate	
  syllable	
  (e.g.,	
  der)	
  than	
  
before	
  the	
  final	
  syllable	
  (e.g.,	
  bee),	
  so	
  that	
  listeners	
  would	
  group	
  the	
  final	
  two	
  
syllables	
  as	
  a	
  single	
  disyllabic	
  word	
  (e.g.,	
  derby).	
  In	
  contrast,	
  the	
  Monosyllabic	
  
context	
  caused	
  the	
  reverse	
  with	
  a	
  stronger	
  prosodic	
  boundary	
  to	
  be	
  heard	
  before	
  
the	
  final	
  syllable	
  (e.g.,	
  bee)	
  than	
  before	
  the	
  penultimate	
  syllable,	
  so	
  that	
  listeners	
  
would	
  hear	
  a	
  final	
  monosyllabic	
  word	
  (e.g.,	
  bee).	
  The	
  Disyllabic	
  context	
  yielded	
  a	
  
(L1-­‐H2-­‐)(L3-­‐H4-­‐)(L5-­‐H6-­‐)(L7-­‐H8)	
  grouping	
  of	
  the	
  final	
  three	
  syllable	
  sequence,	
  
displaying	
  a	
  larger	
  boundary	
  before	
  syllable	
  7	
  than	
  syllable	
  8,	
  thereby	
  critically	
  
yielding	
  a	
  disyllabic	
  final	
  word	
  report	
  (e.g.,	
  derby).	
  In	
  contrast,	
  the	
  Monosyllabic	
  
context	
  yielded	
  a	
  (H1-­‐L2-­‐)(H3-­‐L4-­‐)(HL5-­‐)	
  (H6-­‐L7-­‐)(H8...)	
  grouping	
  of	
  the	
  final	
  three	
  
syllables,	
  displaying	
  a	
  larger	
  boundary	
  before	
  syllable	
  8	
  than	
  before	
  syllable	
  7.	
  -The	
  
pattern	
  of	
  responses	
  by	
  typical	
  listeners	
  was	
  as	
  predicted:	
  there	
  was	
  a	
  strong	
  
tendency	
  for	
  listeners	
  to	
  hear	
  the	
  same	
  acoustic	
  material	
  as	
  ending	
  in	
  disyllabic	
  
words	
  in	
  Disyllabic	
  contexts	
  but	
  as	
  ending	
  in	
  monosyllabic	
  words	
  in	
  Monosyllabic	
  
contexts.	
  Results	
  for	
  these	
  studies	
  showed	
  that	
  typical-­‐developing	
  adults	
  perceived	
  
distal	
  prosodic	
  cues	
  of	
  intonation	
  rhythm	
  in	
  speech	
  and,	
  that	
  their	
  perceptions	
  
manipulated	
  their	
  parsing	
  of	
  syllables	
  in	
  speech	
  perception.	
  	
  
	
  

The	
  studies	
  reported	
  above	
  have	
  provided	
  reliable	
  experiment	
  paradigms	
  for	
  

measuring	
  effects	
  of	
  distal	
  prosodic	
  cues	
  in	
  typical	
  adults.	
  The	
  goal	
  of	
  this	
  thesis	
  is	
  to	
  

	
  

8	
  

extend	
  these	
  paradigms	
  to	
  examine	
  speech	
  perceptions	
  in	
  PWS.	
  These	
  paradigms	
  
require	
  listeners	
  to	
  not	
  only	
  perceive	
  the	
  distal	
  prosodic	
  cues	
  but	
  also,	
  crucially,	
  to	
  
internally	
  generate	
  rhythm	
  and	
  timing	
  from	
  the	
  externally	
  prompted	
  distal	
  prosodic	
  
cues	
  in	
  order	
  to	
  parse	
  the	
  proximal	
  speech	
  material	
  in	
  the	
  predicted	
  way.	
  	
  
	
  

Based	
  on	
  previous	
  literature,	
  it	
  seems	
  plausible	
  that	
  PWS	
  may	
  have	
  a	
  deficit	
  

in	
  their	
  ability	
  to	
  internally	
  generate	
  intonation	
  rhythm	
  and	
  timing	
  used	
  for	
  speech	
  
production.	
  This	
  is	
  supported	
  by	
  the	
  fact	
  that	
  external	
  rhythms,	
  rates,	
  and	
  beats	
  are	
  
evidence-­‐based	
  to	
  help	
  with	
  fluency	
  in	
  this	
  clinical	
  population	
  (Packman	
  et	
  al.,	
  
2007).	
  Studies	
  of	
  the	
  effects	
  of	
  intonation,	
  rhythm,	
  and	
  duration	
  on	
  word	
  
perceptions	
  have	
  never	
  before	
  been	
  conducted	
  with	
  PWS.	
  These	
  studies	
  by	
  Dilley	
  
and	
  Pitt	
  (2010)	
  and	
  Morrill,	
  Dilley,	
  and	
  McAuley	
  (2014)	
  provide	
  an	
  effective	
  way	
  to	
  
measure	
  the	
  influence	
  of	
  prosody	
  in	
  speech	
  perception	
  and	
  allow	
  measuring	
  
sensitivity	
  to	
  prosody	
  by	
  indexing	
  linguistic	
  perception.	
  In	
  the	
  present	
  study,	
  the	
  
sensitivity	
  to	
  prosodic	
  cues	
  was	
  investigated	
  in	
  both	
  a	
  control	
  and	
  clinical	
  
population;	
  this	
  comparison	
  allowed	
  for	
  assessing	
  any	
  group	
  differences	
  and	
  
provided	
  insight	
  into	
  possible	
  perceptual	
  gaps	
  in	
  PWS.	
  	
  Paradigms	
  utilizing	
  
manipulations	
  of	
  distal	
  prosodic	
  cues	
  were	
  utilized	
  due	
  to	
  the	
  demands	
  required	
  by	
  
the	
  listener	
  for	
  distal	
  manipulations	
  compared	
  to	
  proximal	
  manipulations.	
  When	
  the	
  
manipulations	
  of	
  prosodic	
  cues	
  are	
  distal	
  to	
  the	
  lexically	
  ambiguous	
  target	
  portion	
  
of	
  speech,	
  there	
  is	
  a	
  requirement	
  for	
  internally	
  generating	
  the	
  rhythm	
  and	
  timing	
  
signaled	
  distally	
  which	
  prior	
  results	
  suggest	
  should	
  affect	
  perception	
  of	
  speech.	
  The	
  
carry-­‐over	
  of	
  this	
  internally	
  generated	
  rhythm	
  is	
  a	
  valuable	
  measure	
  for	
  

	
  

9	
  

investigation	
  of	
  speech	
  perception	
  within	
  PWS,	
  one	
  that	
  could	
  not	
  be	
  tested	
  using	
  
experimental	
  paradigms	
  that	
  involved	
  proximal	
  prosodic	
  cues.	
  
	
  

This	
  study	
  is	
  motivated	
  by	
  the	
  question	
  of	
  whether	
  people	
  who	
  stutter	
  (PWS)	
  

show	
  a	
  difference	
  in	
  the	
  effects	
  of	
  distal	
  prosodic	
  cues,	
  (pitch	
  rhythms	
  and	
  rate	
  of	
  
speech)	
  on	
  speech	
  perceptions.	
  This	
  question	
  will	
  be	
  investigated	
  from	
  the	
  results	
  of	
  
two	
  experimental	
  paradigms,	
  each	
  utilizing	
  manipulations	
  to	
  different	
  prosodic	
  
characteristics,	
  distally.	
  We	
  hypothesized	
  that	
  PWS	
  would	
  show	
  reduced	
  sensitivity	
  
to	
  the	
  distal	
  rhythm	
  and	
  speech	
  rate	
  manipulations	
  compared	
  with	
  matched	
  
controls.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  	
  
	
  

	
  

10	
  

Participants	
  
Twenty-­‐eight	
  participants	
  (22	
  males;	
  6	
  females),	
  ranging	
  from	
  18-­‐52	
  years	
  old,	
  were	
  
selected	
  for	
  one	
  of	
  two	
  groups:	
  an	
  experimental	
  group	
  of	
  adults	
  who	
  stutter	
  (n=15)	
  
and	
  a	
  control	
  group	
  (n	
  =	
  13).	
  	
  All	
  participants	
  in	
  both	
  groups	
  participated	
  in	
  
Experiments	
  1	
  and	
  2	
  of	
  this	
  study.	
  Participants	
  were	
  right-­‐handed,	
  monolingual	
  
native	
  speakers	
  of	
  American	
  English	
  with	
  normal	
  hearing.	
  	
  Every	
  participant	
  was	
  
without	
  concomitant	
  developmental	
  disorders	
  (e.g.,	
  dyslexia,	
  ADHD,	
  learning	
  delay,	
  
or	
  other	
  confirmed	
  developmental	
  or	
  psychiatric	
  conditions)	
  and	
  was	
  not	
  taking	
  any	
  
medication	
  affecting	
  the	
  central	
  nervous	
  system.	
  Participants	
  were	
  matched	
  on	
  age,	
  
sex,	
  years	
  of	
  education,	
  working	
  memory,	
  receptive	
  vocabulary	
  (Peabody	
  Picture	
  
Vocabulary	
  Test,	
  PPVT-­‐4),	
  expressive	
  vocabulary	
  (Expressive	
  Vocabulary	
  Test,	
  EVT-­‐
2),	
  and	
  a	
  measure	
  of	
  articulation	
  (Goldman-­‐Fristoe	
  Test	
  of	
  Articulation,	
  GFTA-­‐2).	
  	
  
Stuttering	
  severity	
  was	
  assessed	
  off-­‐line	
  by	
  reviewing	
  video	
  recorded	
  samples	
  of	
  
speech	
  elicited	
  through	
  monologue,	
  dialogue,	
  and	
  reading	
  sample	
  with	
  a	
  certified	
  
Speech-­‐Language	
  Pathologist	
  and/or	
  trained	
  Masters	
  student.	
  The	
  Stuttering	
  
Severity	
  Instrument	
  (SSI-­‐4)	
  was	
  used	
  to	
  assess	
  frequency	
  and	
  duration	
  of	
  
stuttering-­‐like	
  and	
  other	
  disfluencies	
  occurring	
  in	
  the	
  speech	
  sample,	
  as	
  well	
  as	
  any	
  
physical	
  concomitants	
  associated	
  with	
  stuttering.	
  These	
  measures	
  were	
  then	
  
incorporated	
  into	
  a	
  composite	
  stuttering	
  severity	
  rating.	
  Participants	
  in	
  the	
  
stuttering	
  group	
  ranged	
  from	
  very	
  mild	
  to	
  severe	
  in	
  severity	
  ratings	
  from	
  the	
  SSI-­‐4.	
  	
  	
  
Table	
  1	
  shows	
  the	
  means	
  and	
  standard	
  deviations	
  for	
  age,	
  handedness,	
  education,	
  
years	
  of	
  musical	
  training,	
  expressive	
  and	
  receptive	
  language	
  measures,	
  articulation	
  

	
  

11	
  

measures,	
  and	
  working	
  memory	
  for	
  both	
  groups,	
  and	
  the	
  stuttering	
  severity	
  for	
  the	
  
stuttering	
  group.	
  
Group	
  
Control	
  
Stuttering	
  
25	
  (7.42)	
  
28.9	
  (11.36)	
  
80.77	
  (17.06)	
  
78	
  (16.12)	
  
15.38	
  (1.39)	
  
16.27	
  (2.25)	
  
5.15	
  (3.89)	
  
3.8	
  (4.71)	
  
107.15	
  (5.98)	
  
109.4	
  (10.2)	
  
107.31	
  (9.6)	
  
115.73	
  (11.32)	
  
100.69	
  (1.44)	
  
100.8	
  (1.66)	
  
40	
  (18.58)	
  
38.8	
  (14.99)	
  

	
  Measure	
  
Age	
  
Handedness	
  
Education	
  
Years	
  of	
  musical	
  training	
  
Peabody	
  Picture	
  Vocabulary	
  Test	
  	
  
Expressive	
  Vocabulary	
  Test	
  	
  
Goldman-­‐Fristoe	
  Test	
  of	
  Articulation	
  
Operation	
  Span	
  Score	
  (Working	
  Memory)	
  

	
  

t-­‐test	
  
p	
  
0.300	
  
0.663	
  
0.232	
  
0.419	
  
0.493	
  
0.045	
  
0.857	
  
0.851	
  

Table	
  1:	
  Participant	
  Information.	
  The	
  table	
  displays	
  statistics	
  of	
  age,	
  handedness,	
  education,	
  
working	
  memory,	
  and	
  speech	
  and	
  language	
  measures.	
  Independent	
  samples	
  t-­‐tests	
  revealed	
  a	
  
significant	
  difference	
  between-­‐subjects	
  on	
  the	
  expressive	
  language	
  measure,	
  in	
  which	
  the	
  stuttering	
  
group	
  displayed	
  a	
  group	
  average	
  standard	
  score	
  8.42	
  points	
  higher	
  than	
  the	
  control	
  group.	
  	
  	
  

	
  
	
  

	
  

	
  

12	
  

Experiment	
  1:	
  Distal	
  Intonation	
  Rhythm	
  
Methods	
  
	
  

The	
  first	
  experiment	
  was	
  a	
  replication	
  and	
  extension	
  of	
  Experiment	
  1A	
  of	
  

Dilley	
  and	
  McAuley	
  (2008).	
  This	
  study	
  utilized	
  distal	
  prosodic	
  manipulations	
  of	
  F0	
  
and	
  duration	
  to	
  examine	
  the	
  effect	
  of	
  distal	
  prosodic	
  cues	
  in	
  linguistic	
  parsing.	
  Eight-­‐
syllable	
  target	
  sequences,	
  such	
  as	
  banker	
  helpful	
  [tɑɪ	
  mɝ˞	
  dɝ	
  bi]	
  were	
  used;	
  the	
  final	
  
four	
  syllables	
  had	
  ambiguous	
  lexical	
  structure	
  and	
  could	
  be	
  organized	
  into	
  words	
  in	
  
more	
  than	
  one	
  way	
  (e.g.,	
  timer	
  derby,	
  tie	
  murder	
  bee).	
  Items	
  selected	
  for	
  this	
  study	
  
corresponded	
  to	
  a	
  subset	
  of	
  items	
  from	
  Dilley,	
  Mattys	
  and	
  Vinke	
  (2010),	
  which	
  were	
  
also	
  used	
  in	
  Morrill	
  et	
  al.	
  (2014).	
  Experimental	
  items	
  utilized	
  one	
  of	
  two	
  different	
  
manipulation	
  patterns	
  of	
  distal	
  F0	
  that	
  was	
  expected	
  to	
  result	
  in	
  strong	
  effects	
  on	
  
the	
  parsing	
  of	
  the	
  final	
  syllables	
  into	
  either	
  a	
  monosyllabic	
  or	
  disyllabic	
  final	
  word.	
  
Monosyllabic	
  and	
  disyllabic	
  contexts	
  were	
  created	
  by	
  varying	
  the	
  fundamental	
  
frequency	
  (FO)	
  across	
  the	
  initial	
  five	
  syllables.	
  The	
  final,	
  target	
  portions	
  of	
  these	
  
sequences	
  remained	
  acoustically	
  ambiguous	
  (containing	
  the	
  same	
  F0	
  pattern)	
  for	
  
both	
  monosyllabic	
  and	
  disyllabic	
  contexts,	
  resulting	
  in	
  the	
  possibility	
  of	
  perceiving	
  
either	
  a	
  monosyllabic	
  word	
  (e.g.,	
  bee)	
  or	
  a	
  disyllabic	
  word	
  (e.g.,	
  derby).	
  	
  
	
  

In	
  one	
  distal	
  prosodic	
  pattern,	
  a	
  repeated	
  low	
  (L)-­‐high	
  (H)	
  alternation	
  of	
  F0	
  

on	
  distal	
  syllables	
  was	
  predicted	
  to	
  lead	
  to	
  grouping	
  of	
  L	
  and	
  H	
  pitch	
  elements	
  
across	
  the	
  entire	
  sequence,	
  which	
  was	
  predicted	
  to	
  generate	
  perception	
  of	
  a	
  
disyllabic	
  word	
  (e.g.,	
  derby);	
  these	
  are	
  referred	
  to	
  as	
  Disyllabic	
  contexts.	
  In	
  the	
  other	
  
distal	
  prosodic	
  pattern,	
  a	
  repeated	
  high-­‐low	
  alternation	
  of	
  F0	
  was	
  predicted	
  to	
  lead	
  
to	
  grouping	
  of	
  H	
  and	
  L	
  pitch	
  elements	
  across	
  the	
  entire	
  sequence,	
  which	
  was	
  
	
  

13	
  

predicted	
  to	
  generate	
  perception	
  of	
  a	
  monosyllabic	
  final	
  word	
  (e.g.,	
  bee);	
  these	
  are	
  
referred	
  to	
  as	
  Monosyllabic	
  contexts.	
  In	
  more	
  detail,	
  for	
  both	
  the	
  monosyllabic	
  and	
  
disyllabic	
  contexts,	
  the	
  final	
  three	
  syllables	
  of	
  each	
  target	
  sequence	
  received	
  a	
  high	
  
(H),	
  low	
  (L),	
  high	
  (H)	
  F0	
  pattern	
  (one	
  F0	
  target	
  per	
  syllable).	
  The	
  first	
  five	
  syllables	
  
of	
  each	
  target	
  sequence	
  varied	
  in	
  F0	
  pattern;	
  the	
  first	
  five	
  syllables	
  of	
  the	
  disyllabic	
  
context	
  items	
  received	
  a	
  L1-­‐H2-­‐L3-­‐H4-­‐L5-­‐	
  pattern,	
  with	
  one	
  F0	
  target,	
  H	
  or	
  L,	
  on	
  
each	
  syllable	
  and	
  the	
  monosyllabic	
  pattern	
  received	
  a	
  H1-­‐L2-­‐H3-­‐L4-­‐	
  HL5-­‐	
  pattern	
  
with	
  one	
  F0	
  target	
  for	
  each	
  of	
  the	
  first	
  four	
  syllables,	
  and	
  a	
  fall	
  in	
  F0	
  from	
  H	
  to	
  L	
  on	
  
the	
  fifth	
  syllable.	
  These	
  distal	
  prosodic	
  context	
  manipulations	
  predicted	
  the	
  
perceptual	
  grouping	
  of	
  the	
  final	
  syllables	
  in	
  the	
  target	
  sequence,	
  in	
  which	
  the	
  
Disyllabic	
  context	
  was	
  predicted	
  to	
  cause	
  a	
  relatively	
  stronger	
  prosodic	
  boundary	
  to	
  
be	
  heard	
  before	
  the	
  penultimate	
  syllable	
  (e.g.,	
  der)	
  than	
  before	
  the	
  final	
  syllable	
  
(e.g.,	
  bee),	
  so	
  that	
  listeners	
  would	
  group	
  the	
  final	
  two	
  syllables	
  as	
  a	
  single	
  disyllabic	
  
word	
  (e.g.,	
  derby).	
  In	
  contrast,	
  the	
  Monosyllabic	
  context	
  was	
  predicted	
  to	
  yield	
  the	
  
reverse,	
  with	
  a	
  stronger	
  prosodic	
  boundary	
  to	
  be	
  heard	
  before	
  the	
  final	
  syllable	
  (e.g.,	
  
bee)	
  than	
  before	
  the	
  penultimate	
  syllable	
  so	
  that	
  listeners	
  would	
  hear	
  a	
  final	
  
monosyllabic	
  word	
  (e.g.,	
  bee).	
  The	
  Disyllabic	
  context	
  yielded	
  a	
  (L1-­‐H2-­‐)(L3-­‐H4-­‐)(L5-­‐
H6-­‐)(L7-­‐H8)	
  grouping	
  of	
  the	
  final	
  three	
  syllable	
  sequence,	
  displaying	
  a	
  larger	
  
boundary	
  before	
  syllable	
  7	
  than	
  syllable	
  8,	
  thereby	
  critically	
  yielding	
  a	
  disyllabic	
  
final	
  word	
  report	
  (e.g.,	
  derby).	
  In	
  contrast,	
  the	
  Monosyllabic	
  context	
  yielded	
  a	
  (H1-­‐
L2-­‐)(H3-­‐L4-­‐)(HL5-­‐)	
  (H6-­‐L7-­‐)(H8)	
  grouping	
  of	
  the	
  final	
  three	
  syllables,	
  displaying	
  a	
  
larger	
  boundary	
  before	
  syllable	
  8	
  than	
  before	
  syllable	
  7.	
  -The	
  pattern	
  of	
  responses	
  
by	
  typical	
  listeners	
  was	
  as	
  predicted:	
  there	
  was	
  a	
  strong	
  tendency	
  for	
  listeners	
  to	
  

	
  

14	
  

hear	
  the	
  same	
  acoustic	
  material	
  as	
  ending	
  in	
  disyllabic	
  words	
  in	
  Disyllabic	
  contexts	
  
but	
  as	
  ending	
  in	
  monosyllabic	
  words	
  in	
  Monosyllabic	
  contexts.	
  .	
  The	
  expected	
  
response	
  for	
  the	
  low-­‐high	
  intonation	
  pattern	
  of	
  the	
  experimental	
  items	
  is	
  to	
  
perceive	
  and	
  report	
  a	
  disyllabic	
  word	
  (derby)	
  as	
  long	
  as	
  the	
  listener	
  is	
  perceiving	
  
the	
  externally	
  cued	
  distal	
  intonation	
  rhythm	
  and	
  internally	
  carrying	
  it	
  throughout	
  
the	
  final	
  two	
  target	
  syllables.	
  These	
  results	
  were	
  expected	
  to	
  be	
  seen	
  within	
  the	
  
control	
  group,	
  and	
  seen	
  as	
  a	
  higher	
  difference	
  between	
  the	
  reported	
  number	
  od	
  
disyllabic	
  words	
  and	
  monosyllabic	
  words	
  for	
  either	
  disyllabic	
  or	
  monosyllabic	
  
contexts.	
  We	
  expected	
  PWS	
  to	
  demonstrate	
  a	
  lesser	
  difference	
  between	
  the	
  
proportions	
  of	
  reported	
  disyllabic	
  and	
  monosyllabic	
  final	
  words.	
  	
  
	
  

Twenty	
  experimental	
  stimuli	
  were	
  provided	
  aurally	
  to	
  each	
  participant	
  via	
  

headphones,	
  with	
  each	
  auditory	
  speech	
  sequence	
  containing	
  two	
  disyllabic,	
  non-­‐
compound	
  words	
  continued	
  with	
  a	
  string	
  of	
  four	
  syllables	
  [tɑɪ	
  mɝ˞ dɝ˞ bi].	
  Both	
  
target	
  and	
  filler	
  sentences	
  were	
  read	
  as	
  connected	
  speech,	
  using	
  monotone	
  F0.	
  
Resynthesized	
  speech	
  stimuli	
  were	
  then	
  derived	
  from	
  these	
  utterances	
  using	
  the	
  
pitch-­‐synchronous	
  overlap-­‐and-­‐add	
  (PSOLA)	
  algorithm	
  (Moulines	
  &	
  Charpentier,	
  
1990)	
  as	
  implemented	
  in	
  Praat	
  software.	
  	
  Methods	
  for	
  recording	
  are	
  described	
  in	
  
Dilley,	
  Mattys	
  and	
  Vinke	
  (2010).	
  	
  
Ten	
  filler	
  sentences	
  were	
  included	
  in	
  the	
  experiment	
  to	
  serve	
  as	
  practice	
  and	
  
disguise	
  the	
  lexical	
  ambiguity	
  present	
  in	
  target	
  sentences.	
  	
  Fillers	
  consisted	
  of	
  a	
  
mixture	
  of	
  monosyllabic	
  and	
  disyllabic	
  words	
  with	
  unambiguous	
  lexical	
  structure	
  in	
  
varying	
  positions	
  within	
  the	
  string;	
  five	
  of	
  the	
  filler	
  sequences	
  ended	
  in	
  a	
  disyllabic	
  
word	
  and	
  the	
  other	
  five	
  ended	
  in	
  a	
  monosyllabic	
  word.	
  Each	
  experimental	
  item	
  
	
  

15	
  

consisted	
  of	
  initial	
  primary	
  stress	
  on	
  the	
  two	
  disyllabic	
  word	
  sequence,	
  e.g.,	
  channel	
  
dizzy,	
  followed	
  by	
  a	
  four-­‐syllable	
  sequence	
  that	
  could	
  be	
  grouped	
  into	
  words	
  in	
  
more	
  than	
  one	
  way.	
  

	
  

Figure	
  1:	
  Distal	
  Context	
  Manipulations.	
  Example	
  of	
  the	
  distal	
  context	
  intonation	
  pattern	
  for	
  stimuli	
  
used	
  in	
  this	
  experiment.	
  The	
  top	
  row	
  displays	
  a	
  sample	
  speech	
  waveform	
  of	
  an	
  auditory	
  string.	
  The	
  
bottom	
  two	
  rows	
  show	
  the	
  fundamental	
  frequency	
  (F0)	
  contour	
  for	
  the	
  two	
  different	
  contexts	
  
(monosyllabic	
  vs.	
  disyllabic)	
  along	
  with	
  the	
  expected	
  syllable	
  parsing	
  for	
  each.	
  H	
  and	
  L	
  refer	
  to	
  high	
  
and	
  low	
  F0	
  targets,	
  respectively.	
  The	
  middle	
  row	
  displays	
  the	
  intonation	
  sequence	
  and	
  expected	
  
syllable	
  parsing	
  for	
  the	
  monosyllabic	
  context,	
  and	
  the	
  bottom	
  for	
  row	
  for	
  the	
  disyllabic	
  context.	
  Note	
  
the	
  targeted	
  final	
  two	
  syllables	
  for	
  both	
  contexts	
  remain	
  the	
  same	
  low-­‐high	
  pattern,	
  and	
  the	
  lexical	
  
content	
  remaining	
  ambiguous	
  with	
  logical	
  parsing	
  for	
  either	
  a	
  monosyllabic	
  word	
  (still)	
  or	
  a	
  
disyllabic	
  word	
  (standstill).	
  	
  
	
  

The	
  distal	
  prosodic	
  context	
  was	
  expected	
  to	
  affect	
  participants’	
  proportions	
  of	
  
disyllabic	
  final	
  words	
  heard,	
  although	
  the	
  portion	
  of	
  the	
  string	
  being	
  judged	
  
consisted	
  of	
  identical	
  acoustic	
  material	
  across	
  conditions.	
  	
  
	
  	
  

Participants	
  provided	
  a	
  free	
  recall	
  typed	
  report	
  of	
  the	
  final	
  word	
  they	
  heard	
  

for	
  each	
  sentence,	
  with	
  the	
  dependent	
  measure	
  being	
  the	
  proportion	
  of	
  disyllabic	
  
responses	
  reported	
  for	
  each	
  experimental	
  stimulus.	
  Each	
  participant	
  parsed	
  the	
  
final	
  four	
  syllables	
  of	
  the	
  string	
  into	
  either	
  a	
  disyllabic	
  final	
  word	
  (e.g.,	
  timer	
  derby)	
  
or	
  a	
  monosyllabic	
  final	
  word	
  (e.g.,	
  tie	
  murder	
  bee).	
  All	
  experimental	
  items	
  were	
  

	
  

16	
  

presented	
  in	
  both	
  a	
  monosyllabic	
  and	
  disyllabic	
  context,	
  and	
  were	
  counterbalanced	
  
across	
  participants.	
  The	
  entire	
  experiment	
  took	
  about	
  12	
  minutes	
  to	
  complete.	
  
Results	
  
Measures	
  were	
  taken	
  on	
  the	
  proportion	
  of	
  reported	
  disyllabic	
  final	
  words	
  as	
  
a	
  function	
  of	
  type	
  of	
  word	
  context	
  (disyllabic	
  vs.	
  monosyllabic)	
  and	
  group	
  (control	
  
vs.	
  stuttering).	
  The	
  typed	
  reported	
  words	
  for	
  each	
  participant	
  were	
  coded	
  by	
  two	
  
different	
  raters	
  as	
  being	
  either	
  a	
  monosyllabic	
  or	
  disyllabic	
  word.	
  There	
  were	
  no	
  
discrepancies	
  between	
  the	
  two	
  raters,	
  and	
  every	
  participant’s	
  responses	
  were	
  
included	
  in	
  the	
  analysis.	
  	
  
Table	
  2	
  displays	
  the	
  means	
  and	
  standard	
  deviations	
  for	
  both	
  the	
  disyllabic	
  
and	
  monosyllabic	
  words	
  reported	
  across	
  the	
  groups.	
  A	
  2	
  (type	
  of	
  word	
  context:	
  
monosyllabic	
  vs.	
  disyllabic)	
  x	
  2	
  (group:	
  control	
  vs.	
  stuttering)	
  mixed-­‐measures	
  
ANOVA	
  on	
  disyllabic	
  response	
  proportions	
  was	
  conducted.	
  First,	
  there	
  was	
  a	
  
significant	
  main	
  effect	
  of	
  Word	
  Context	
  (F(1,	
  26)	
  =	
  337.121,	
  p	
  <	
  .001,	
  η p2 =	
  .928),	
  	
  
corresponding	
  to	
  a	
  higher	
  proportion	
  of	
  disyllabic	
  responses	
  in	
  the	
  disyllabic	
  
context	
  (M	
  =	
  .940)	
  than	
  the	
  monosyllabic	
  context	
  (M	
  =	
  .164).	
  	
  These	
  results	
  are	
  
consistent	
  with	
  the	
  expected	
  higher	
  proportion	
  of	
  disyllabic	
  responses	
  due	
  to	
  effects	
  
of	
  Word	
  Context.	
  

	
  

17	
  

Proportion	
  of	
  Disyllabic	
  Responses	
  

1	
  
0.9	
  
0.8	
  
0.7	
  
0.6	
  
0.5	
  

Stuttering	
  

0.4	
  

Control	
  

0.3	
  
0.2	
  
0.1	
  
0	
  
Di	
  

Mono	
  
Word	
  Context	
  

	
  
Figure	
  2:	
  Proportion	
  of	
  Disyllabic	
  Responses.	
  The	
  effects	
  of	
  Word	
  Context	
  within-­‐subjects	
  on	
  
proportion	
  of	
  reported	
  disyllabic	
  responses	
  for	
  both	
  experimental	
  stimuli	
  contexts,	
  is	
  displayed	
  
above.	
  Stuttering	
  refers	
  to	
  the	
  experimental	
  group	
  of	
  adults	
  who	
  stutter,	
  and	
  control	
  refers	
  to	
  group	
  
of	
  typically-­‐fluent	
  speakers.	
  

	
  
No	
  significant	
  effect	
  of	
  group	
  (F(1,	
  26)	
  =	
  1.384,	
  p	
  =	
  .250,	
  η p2 	
  =	
  .051)	
  or	
  interaction	
  
(F(1,	
  26)	
  =	
  1.720,	
  p	
  =	
  .201,	
  η p2 	
  =	
  .062)	
  was	
  found.	
  
	
  
Discussion	
  
	
  

In	
  Experiment	
  1,	
  we	
  examined	
  how	
  the	
  stuttering	
  group	
  would	
  respond	
  to	
  

the	
  manipulation	
  of	
  the	
  distal	
  prosodic	
  cues	
  in	
  lexical	
  segmentation	
  of	
  an	
  ambiguous	
  
target	
  sequence	
  by	
  measuring	
  the	
  proportion	
  of	
  reported	
  disyllabic	
  final	
  words	
  for	
  
each	
  experimental	
  stimulus.	
  It	
  was	
  hypothesized	
  that	
  PWS	
  would	
  be	
  less	
  sensitive	
  to	
  
this	
  rhythmic	
  cuing,	
  resulting	
  in	
  a	
  smaller	
  difference	
  in	
  the	
  proportion	
  of	
  disyllabic	
  
words	
  in	
  Disyllabic	
  vs.	
  Monosyllabic	
  distal	
  contexts	
  compared	
  with	
  control	
  
participants.	
  Our	
  finding	
  of	
  no	
  significant	
  effect	
  of	
  group	
  demonstrates	
  that	
  PWS	
  

	
  

18	
  

showed	
  similar	
  sensitivity	
  to	
  rhythmic	
  prosody	
  compared	
  with	
  the	
  control	
  group.	
  	
  
The	
  significant	
  main	
  effect	
  of	
  Word	
  Context	
  within-­‐subject	
  for	
  both	
  groups	
  indicates	
  
that	
  both	
  PWS	
  and	
  the	
  controls	
  exhibited	
  sensitivity	
  to	
  distal	
  prosodic	
  cues,	
  which	
  
influenced	
  their	
  syllabic	
  parsing	
  of	
  the	
  final	
  target	
  string.	
  	
  The	
  design	
  of	
  this	
  
experimental	
  paradigm	
  relies	
  on	
  the	
  sensitivity	
  of	
  the	
  listener	
  to	
  not	
  only	
  perceive	
  
the	
  distal	
  prosodic	
  cues	
  of	
  manipulated	
  F0	
  and	
  duration,	
  but	
  also	
  internally	
  generate	
  
the	
  intonation	
  rhythm	
  and	
  carry	
  it	
  through	
  to	
  the	
  final	
  four	
  syllables	
  in	
  order	
  to	
  
parse	
  a	
  disyllabic	
  final	
  word.	
  	
  The	
  PWS	
  group	
  resulted	
  in	
  reporting	
  a	
  higher	
  
proportion	
  of	
  disyllabic	
  words	
  than	
  monosyllabic	
  words,	
  similar	
  to	
  pattern	
  
exhibited	
  by	
  the	
  control	
  group.	
  Thus,	
  we	
  can	
  conclude	
  that	
  PWS	
  have	
  similar	
  
sensitivity	
  to	
  prosodic	
  elements	
  in	
  their	
  speech	
  perceptions,	
  as	
  indexed	
  by	
  these	
  
distal	
  prosody	
  manipulations,	
  as	
  compared	
  with	
  fluent	
  speakers.	
  These	
  results	
  also	
  
suggest	
  that	
  PWS	
  were	
  able	
  to	
  internally	
  generate	
  the	
  intonation	
  rhythm	
  perceived	
  
throughout	
  the	
  distal	
  material	
  in	
  a	
  manner	
  that	
  carried	
  over	
  to	
  the	
  lexical	
  
ambiguous	
  targeted	
  sequence;	
  this	
  suggests	
  that	
  under	
  the	
  present	
  experimental	
  
conditions	
  the	
  internal	
  rhythm	
  generation	
  of	
  PWS	
  was	
  intact	
  for	
  a	
  minimal	
  time-­‐
span	
  of	
  carryover.	
  Therefore,	
  we	
  did	
  not	
  find	
  support	
  for	
  the	
  hypothesis	
  that	
  PWS	
  
would	
  demonstrate	
  a	
  lessened	
  sensitivity	
  to	
  the	
  distal	
  prosodic	
  cuing	
  of	
  intonation	
  
rhythm	
  and	
  duration.	
  	
  
	
  

	
  

	
  

19	
  

Experiment	
  2:	
  Distal	
  Speech	
  Rate	
  
Methods	
  
The	
  second	
  experiment	
  used	
  in	
  this	
  study	
  was	
  a	
  replication	
  and	
  
extension	
  of	
  the	
  experimental	
  paradigm	
  used	
  by	
  Dilley	
  and	
  Pitt	
  (2010).	
  This	
  study	
  
investigated	
  sensitivity	
  to	
  prosodic	
  cues	
  at	
  the	
  syllable	
  level	
  by	
  manipulating	
  
context	
  speech	
  rate.	
  The	
  experiment	
  utilized	
  40	
  experimental	
  items;	
  each	
  included	
  a	
  
sentence	
  that	
  contained	
  target	
  syllables	
  with	
  a	
  function	
  word	
  presented	
  via	
  auditory	
  
recordings.	
  Each	
  sentence	
  had	
  a	
  function	
  word	
  that	
  is	
  optional,	
  as	
  the	
  sentence	
  will	
  
be	
  grammatical	
  with	
  or	
  without	
  the	
  function	
  word.	
  For	
  example,	
  in	
  the	
  sentence	
  
“Don	
  must	
  see	
  the	
  harbor	
  (or)	
  boats,”	
  the	
  target	
  region	
  consisted	
  of	
  the	
  function	
  
word	
  or,	
  the	
  preceding	
  syllable	
  –bor,	
  and	
  the	
  following	
  phoneme	
  [b].	
  The	
  context	
  
region	
  included	
  the	
  non-­‐target	
  portions	
  of	
  the	
  context.	
  

	
  
Figure	
  3:	
  Distal	
  Rate	
  Manipulations.	
  Waveforms	
  of	
  a	
  sample	
  altered	
  stimulus	
  exampling	
  normal	
  
(unaltered	
  rate)	
  versus	
  a	
  slowed	
  distal	
  context	
  (slowed	
  rate).	
  Sections	
  of	
  the	
  waveform	
  without	
  
background	
  shadowing	
  correspond	
  to	
  the	
  target	
  region,	
  which	
  consisted	
  of	
  a	
  critical	
  function	
  word	
  
(e.g.,	
  “or”)	
  plus	
  the	
  preceding	
  syllable	
  and	
  following	
  phoneme,	
  following	
  Dilley	
  and	
  Pitt	
  (2010).	
  
	
  

	
  

The	
  experimental	
  stimuli	
  used	
  four	
  different	
  speech	
  rate	
  expansion	
  factors	
  
(1,	
  1.3,	
  1.6,	
  1.9),	
  each	
  involving	
  a	
  temporal	
  manipulation	
  in	
  which	
  the	
  distal	
  context	
  
of	
  the	
  phrase	
  is	
  modified	
  while	
  the	
  target	
  syllables	
  remain	
  un-­‐modified.	
  For	
  slowed	
  

	
  

20	
  

rates,	
  both	
  the	
  portion	
  of	
  speech	
  preceding	
  the	
  target	
  region	
  and	
  the	
  region	
  that	
  
followed	
  it	
  were	
  multiplied	
  by	
  1.3,	
  1.6,	
  or	
  1.9	
  using	
  the	
  PSOLA	
  algorithm	
  in	
  Praat	
  
(Dilley	
  &	
  Pitt,	
  2010).	
  That	
  is,	
  the	
  duration	
  of	
  the	
  context	
  was	
  130%,	
  160%,	
  or	
  190%	
  
the	
  duration	
  of	
  the	
  original	
  context,	
  thus	
  slowing	
  the	
  speech	
  rate.	
  The	
  unaltered	
  
items	
  were	
  multiplied	
  by	
  1.0	
  using	
  the	
  PSOLA	
  algorithm,	
  maintaining	
  their	
  original	
  
speech	
  rate.	
  It	
  was	
  expected	
  that	
  if	
  the	
  listener	
  had	
  sensitivity	
  to	
  the	
  distal	
  speech	
  
rate	
  cues,	
  he	
  or	
  she	
  would	
  report	
  a	
  successively	
  smaller	
  proportion	
  of	
  function	
  
words	
  as	
  the	
  distal	
  speech	
  rate	
  was	
  made	
  increasingly	
  slow.	
  The	
  dependent	
  
measure	
  was	
  the	
  proportion	
  of	
  reported	
  function	
  words.	
  We	
  predicted	
  that	
  if	
  PWS	
  
exhibited	
  a	
  lessened	
  sensitivity	
  to	
  distal	
  speech	
  rate	
  compared	
  with	
  control	
  
participants,	
  there	
  would	
  be	
  a	
  smaller	
  drop	
  in	
  the	
  proportion	
  of	
  function	
  words	
  with	
  
increasingly	
  slow	
  speech	
  rates	
  compared	
  with	
  controls.	
  Each	
  participant	
  heard	
  
every	
  experimental	
  sentence	
  and	
  the	
  pairing	
  of	
  speech	
  rates	
  with	
  experimental	
  
items	
  was	
  counterbalanced	
  across	
  participants.	
  Filler	
  items	
  had	
  no	
  ambiguity	
  
regarding the number of words they contained. They were presented at either the
unaltered or slowed rate.
	
  

The	
  participants	
  were	
  shown	
  the	
  first	
  four	
  words	
  of	
  the	
  phrase	
  on	
  the	
  

computer	
  screen	
  and	
  were	
  asked	
  to	
  transcribe	
  the	
  remainder	
  of	
  the	
  words	
  they	
  
heard.	
  If	
  the	
  participant	
  showed	
  sensitivity	
  to	
  the	
  distal	
  speech	
  rate	
  manipulations,	
  
he/she	
  did	
  not	
  perceive	
  the	
  function	
  word	
  when	
  the	
  distal	
  context	
  was	
  slowed,	
  but	
  
did	
  perceive	
  the	
  function	
  word	
  when	
  the	
  distal	
  context	
  was	
  un-­‐modified.	
  	
  Each	
  
participant’s	
  responses	
  were	
  coded	
  by	
  two	
  different	
  raters.	
  The	
  reported	
  function	
  
word	
  was	
  determined	
  by	
  the	
  following	
  definition;	
  a	
  word	
  (such	
  as	
  a	
  preposition	
  or	
  a	
  

	
  

21	
  

conjunction)	
  that	
  is	
  used	
  mainly	
  to	
  show	
  grammatical	
  relationships	
  between	
  other	
  
words.	
  If	
  the	
  reported	
  function	
  word	
  was	
  not	
  the	
  same	
  as	
  the	
  spoken	
  function	
  word	
  
in	
  the	
  auditory	
  stimuli,	
  it	
  was	
  still	
  counted	
  as	
  a	
  reported	
  function	
  word	
  with	
  the	
  
constraints	
  that	
  the	
  immediate	
  preceding	
  and	
  proceeding	
  phonemic	
  environments	
  
of	
  the	
  target	
  context	
  remained	
  accurate.	
  
Results	
  	
  
Analyses were based on whether participants reported hearing a function word in
the acoustically ambiguous region for each utterance. The frequency of transcribing a
function word in the target region was scored; presence of a critical function word in the
target region was coded as 1, and absence of a function word was coded as 0. Responses
that indicated inattention to the stimuli (i.e., did not include the preceding and/or
following phonemic environments) were not included in the analysis (9%).
Figure 4 displays the proportion of reported function words as a function of rate
and group. A 4 (Rate: 1, 1.3, 1.6, 1.9) x 2 (Group: control vs. stuttering) mixed-measures
ANOVA on reported function word was conducted. There was a significant main effect
of rate (F(3,	
  78)	
  =	
  118.230,	
  p	
  <	
  .001,	
  η p2 =	
  .820).	
  

	
  

22	
  

Average	
  Reported	
  Function	
  Word	
  

1	
  
0.9	
  
0.8	
  
0.7	
  
0.6	
  
0.5	
  
0.4	
  
0.3	
  
0.2	
  
0.1	
  
0	
  

Stuttering	
  
Control	
  

1	
  

1.3	
  

1.6	
  

1.9	
  

Rate	
  
Figure	
  4:	
  Reported	
  Function	
  Word	
  by	
  Rate.	
  Average	
  reported	
  function	
  word	
  for	
  each	
  rate	
  
between-­‐subjects	
  displayed	
  no	
  significant	
  differences	
  between	
  stuttering	
  and	
  control	
  groups.	
  	
  Distal	
  
context	
  rate	
  had	
  a	
  significant	
  main	
  effect	
  for	
  both	
  groups	
  resulting	
  in	
  a	
  gradual	
  decline	
  of	
  average	
  
reported	
  function	
  words	
  as	
  the	
  rate	
  was	
  slowed.	
  	
  

There was no significant main effect of group (F(1,	
  26)	
  =	
  .032,	
  p	
  =	
  .860,	
  η p2 	
  =	
  .001) or	
  
interaction	
  (F(1,	
  26)	
  =	
  262.028,	
  p	
  =	
  .820,	
  η p2 	
  =	
  .019)	
  found.	
  Given	
  the	
  significant	
  
effect	
  of	
  Rate,	
  Bonferroni-­‐corrected,	
  post-­‐hoc	
  t-­‐tests	
  were	
  conducted	
  to	
  further	
  
examine	
  pairwise	
  differences	
  among	
  the	
  levels. Results demonstrated significant
differences between all pairs of rate levels at p	
  < .001 except for 1.6 vs. 1.9 (p	
  = .057).
These results demonstrate that the speech rate effect is robust across both PWS and
control participants.
A few additional observations are noteworthy about this data. First, in the normalrate condition, function word reports were quite high; the fact that they were not at
ceiling is expected given that the speech was casually spoken. Moreover, note that a
comparison of reports in the normal-rate condition and the slowed-context conditions
(1.3, 1.6, and 1.9) showed that merely slowing the context surrounding a function word
caused the rate of reported function words to drop. The degree of drop was shown to be

	
  

23	
  

consistent across both groups. These results display a gradual decrease in average
reported function word as a function of slowed distal rate outside of the final two rates, in
which the decrease of the reported function word means was not significant.
Discussion	
  
	
  

Experiment	
  2	
  showed	
  a	
  significant	
  main	
  effect	
  of	
  distal	
  speech	
  rate,	
  but	
  PWS	
  

and	
  control	
  participants	
  showed	
  similar	
  sensitivity	
  to	
  distal	
  speech	
  rate,	
  resulting	
  in	
  
no	
  significant	
  effect	
  of	
  group.	
  These	
  results	
  did	
  not	
  support	
  our	
  original	
  hypothesis	
  
that	
  PWS	
  would	
  exhibit	
  a	
  lessened	
  sensitivity	
  in	
  perception	
  of	
  distal	
  speech	
  rate,	
  and	
  
therefore	
  would	
  not	
  have	
  as	
  large	
  of	
  a	
  difference	
  of	
  reported	
  frequency	
  words	
  
between	
  distal	
  speech	
  rates	
  as	
  the	
  control	
  population	
  would.	
  Making	
  the	
  duration	
  of	
  
a	
  stretch	
  of	
  speech	
  containing	
  a	
  function	
  word	
  fast	
  relative	
  to	
  its	
  slowed	
  distal	
  
context	
  affected	
  the	
  number	
  of	
  morphophonological	
  units	
  perceived	
  by	
  the	
  
participants	
  in	
  both	
  groups	
  to	
  a	
  similar	
  degree.	
  We	
  obtained	
  clear	
  evidence	
  that	
  an	
  
alteration	
  of	
  speech	
  rate	
  on	
  the	
  distal	
  context	
  affected	
  the	
  word	
  segmentation	
  of	
  the	
  
target	
  context,	
  extending	
  prior	
  work	
  by	
  Dilley	
  and	
  Pitt	
  (2010)	
  to	
  a	
  clinical	
  group	
  of	
  
PWS.	
  Due	
  to	
  the	
  lack	
  of	
  group	
  differences	
  in	
  the	
  comparison	
  of	
  PWS	
  and	
  controls,	
  we	
  
can	
  infer	
  that	
  PWS	
  exhibited	
  sensitivity	
  to	
  distal	
  speech	
  rate	
  and	
  did	
  not	
  
demonstrate	
  impairment	
  in	
  their	
  perception	
  or	
  internal	
  generation	
  of	
  the	
  distal	
  rate	
  
throughout	
  the	
  stimulus.	
  	
  
	
  

	
  

	
  

24	
  

General	
  Discussion	
  and	
  Conclusions	
  
Stuttering	
  is	
  characterized	
  as	
  a	
  neurodevelopmental	
  disorder	
  that	
  impacts	
  
the	
  rhythmic	
  flow	
  and	
  timing	
  of	
  spoken	
  language.	
  	
  Despite	
  the	
  amount	
  of	
  previous	
  
research	
  and	
  plausible	
  theories	
  on	
  the	
  etiology	
  of	
  stuttering,	
  there	
  still	
  is	
  no	
  
generally	
  accepted	
  cause	
  of	
  the	
  underlying	
  mechanisms	
  affecting	
  fluency.	
  	
  Prior	
  
evidence	
  has	
  shown	
  that	
  fluency-­‐enhancing	
  techniques	
  for	
  PWS	
  utilize	
  some	
  type	
  of	
  
externally-­‐generated	
  rhythm	
  or	
  timing	
  including	
  singing,	
  choral	
  reading,	
  or	
  using	
  a	
  
metronome,	
  suggesting	
  that	
  internal	
  beat	
  and	
  rhythm	
  generation	
  is	
  problematic	
  for	
  
PWS	
  (Adams	
  &	
  Ramig,	
  1980;	
  Ingham	
  &	
  Carroll,	
  1977;	
  Glover,	
  Kalinowski,	
  Rastatter,	
  
&	
  Stuart,	
  1996;	
  Wingate	
  &	
  Howell,	
  2002).	
  	
  Other	
  research	
  supported	
  this	
  idea	
  with	
  
evidence	
  that	
  children	
  who	
  stutter	
  displayed	
  lessened	
  white	
  matter	
  activation	
  in	
  the	
  
basal	
  ganglia	
  thalamocortical	
  (BGTC)	
  network,	
  which	
  involves	
  the	
  neural	
  cortices	
  
that	
  are	
  utilized	
  in	
  the	
  auditory	
  perceptions	
  of	
  rhythm,	
  beat,	
  and	
  timing	
  (Chang	
  &	
  
Zhu,	
  2013;	
  Grahn	
  &	
  McAuley,	
  2009;Toyomura	
  et	
  al.,	
  2015).	
  While	
  stuttering	
  is	
  
typically	
  thought	
  of	
  a	
  speech	
  production	
  disorder,	
  studies	
  have	
  supported	
  that	
  
speech	
  productions	
  of	
  people	
  who	
  stutter	
  (PWS)	
  can	
  be	
  influenced	
  by	
  the	
  
perception	
  of	
  prosodic	
  aspects	
  of	
  their	
  own	
  speech	
  (Hargrave,	
  Kalinowski,	
  Stuart,	
  
Armson,	
  &	
  Jones,	
  1994;	
  Wingate,	
  1966).	
  	
  This	
  study	
  was	
  thus	
  conducted	
  to	
  better	
  
understand	
  the	
  nature	
  of	
  the	
  possible	
  deficits	
  in	
  the	
  perceptions	
  of	
  rhythm	
  and	
  
timing	
  and	
  internal	
  beat	
  generation	
  in	
  persons	
  who	
  stutter.	
  	
  We	
  investigated	
  the	
  
perception	
  of	
  distal	
  speech	
  rhythm	
  and	
  rate	
  in	
  a	
  group	
  of	
  adults	
  who	
  stutter	
  and	
  a	
  
control	
  group.	
  	
  	
  

	
  

25	
  

In	
  Experiment	
  1,	
  distal	
  intonation	
  rhythm	
  was	
  manipulated	
  and	
  both	
  groups	
  
responded	
  with	
  the	
  final	
  word	
  that	
  they	
  heard.	
  The	
  experimental	
  manipulations	
  
utilized	
  distal	
  pitch	
  pattern	
  which	
  were	
  expected	
  to	
  cause	
  differential	
  perception	
  of	
  
disyllabic	
  vs.	
  monosyllabic	
  final	
  words	
  if	
  they	
  were	
  sensitive	
  to	
  distal	
  cues	
  and	
  could	
  
use	
  them	
  to	
  generate	
  a	
  rhythm	
  (Dilley	
  &	
  McAuley,	
  2008).	
  	
  Our	
  hypothesis	
  proposed	
  
that	
  the	
  stuttering	
  group	
  would	
  show	
  less	
  sensitivity	
  to	
  these	
  prosodic	
  
manipulations,	
  which	
  was	
  predicted	
  to	
  manifest	
  as	
  a	
  smaller	
  difference	
  in	
  disyllabic	
  
final	
  word	
  reports	
  in	
  disyllabic	
  vs.	
  monosyllabic	
  context	
  conditions.	
  	
  In	
  Experiment	
  
1,	
  PWS	
  reported	
  a	
  higher	
  proportion	
  of	
  disyllabic	
  responses	
  in	
  disyllabic	
  contexts	
  
and	
  similar	
  effects	
  of	
  context	
  type	
  as	
  the	
  control	
  group,	
  indicating	
  that	
  they	
  
perceived	
  the	
  distal	
  prosodic	
  manipulations	
  and	
  internally	
  generated	
  a	
  rhythm	
  that	
  
influenced	
  the	
  syllable	
  parsing	
  of	
  the	
  final	
  word	
  in	
  a	
  manner	
  that	
  was	
  similar	
  to	
  
controls.	
  	
  	
  
In	
  Experiment	
  2,	
  distal	
  rate	
  of	
  speech	
  was	
  manipulated	
  and	
  both	
  groups	
  
reported	
  the	
  final	
  words,	
  with	
  the	
  dependent	
  variable	
  being	
  function	
  word	
  reports.	
  	
  
The	
  experimental	
  manipulations	
  utilized	
  a	
  slowing	
  of	
  distal	
  speech	
  rate	
  to	
  cause	
  the	
  
function	
  word	
  to	
  be	
  heard	
  increasingly	
  less	
  often	
  with	
  increasingly	
  slow	
  distal	
  
speech	
  rates.	
  	
  Our	
  hypothesis	
  proposed	
  that	
  PWS	
  would	
  show	
  a	
  lessened	
  sensitivity	
  
to	
  the	
  distal	
  prosodic	
  cues,	
  resulting	
  in	
  a	
  smaller	
  decrease	
  in	
  function	
  word	
  reports	
  
as	
  distal	
  speech	
  rate	
  was	
  slowed,	
  compared	
  with	
  controls.	
  In	
  Experiment	
  2,	
  the	
  
significant	
  main	
  effect	
  of	
  rate	
  in	
  PWS	
  indicated	
  that	
  PWS	
  were	
  sensitive	
  to	
  the	
  distal	
  
manipulations	
  of	
  rate,	
  since	
  they	
  reported	
  fewer	
  function	
  words	
  as	
  distal	
  was	
  
slowed,	
  mirroring	
  the	
  pattern	
  for	
  controls.	
  	
  

	
  

26	
  

	
  	
  

Based	
  on	
  our	
  hypothesis,	
  if	
  PWS	
  have	
  a	
  deficit	
  with	
  internally	
  generating	
  

rhythm	
  and	
  timing,	
  we	
  might	
  expect	
  that	
  they	
  would	
  have	
  difficulty	
  generating	
  an	
  
internal	
  structure	
  from	
  perceived	
  speech	
  rhythmic	
  cues	
  that	
  would	
  manifest	
  as	
  a	
  
difference	
  between	
  control	
  and	
  PWS	
  in	
  a	
  task	
  of	
  distal	
  prosody.	
  	
  Both	
  Experiments	
  1	
  
and	
  2	
  showed	
  that	
  PWS	
  performed	
  similarly	
  in	
  these	
  tasks	
  to	
  the	
  control	
  group.	
  In	
  
Experiment	
  1,	
  the	
  significant	
  main	
  effect	
  of	
  word	
  context	
  for	
  both	
  groups	
  with	
  no	
  
significant	
  effect	
  of	
  group	
  stood	
  in	
  contrast	
  to	
  the	
  hypothesis	
  of	
  lessened	
  sensitivity	
  
to	
  the	
  distal	
  prosodic	
  manipulations	
  of	
  intonation	
  rhythm	
  in	
  PWS.	
  	
  
These	
  results	
  are	
  inconsistent	
  with	
  the	
  idea	
  that	
  the	
  ability	
  to	
  generate	
  
prosody	
  perceptually	
  is	
  not	
  intact	
  in	
  PWS.	
  That	
  is,	
  PWS	
  did	
  not	
  demonstrate	
  an	
  
impaired	
  function	
  within	
  their	
  perceptions	
  of	
  rhythm	
  and	
  timing	
  cues	
  in	
  distal	
  
speech	
  or	
  in	
  their	
  ability	
  to	
  internally	
  generate	
  rhythms	
  and	
  rates	
  cued	
  externally	
  
throughout	
  the	
  ambiguous	
  and	
  unaltered	
  targeted	
  speech	
  strings.	
  	
  This	
  conclusion	
  
on	
  the	
  surface	
  appears	
  inconsistent	
  with	
  proposals	
  of	
  deficits	
  in	
  rhythm	
  perception	
  
caused	
  from	
  dysfunction	
  within	
  the	
  basal	
  ganglia	
  and	
  the	
  BGTC	
  network	
  (Grahn,	
  
2009).	
  Evidence	
  of	
  a	
  deficit	
  in	
  rhythm	
  perception	
  was	
  recently	
  found	
  by	
  Wieland	
  
(2015)	
  in	
  children	
  who	
  stutter	
  (CWS).	
  	
  This	
  evidence	
  is	
  also	
  supported	
  by	
  the	
  
reduced	
  white	
  matter	
  tracts	
  found	
  in	
  CWS	
  by	
  Chang	
  (2013)	
  compared	
  with	
  controls;	
  
these	
  tracts	
  are	
  expected	
  to	
  affect	
  connectivity	
  in	
  the	
  BGTC	
  network.	
  	
  
The	
  failure	
  to	
  find	
  evidence	
  for	
  deficits	
  in	
  perception	
  of	
  rhythmic	
  aspects	
  of	
  
speech	
  and	
  apparent	
  conflict	
  with	
  findings	
  from	
  previous	
  literature	
  could	
  be	
  
explained	
  with	
  reference	
  to	
  four	
  proposals.	
  	
  The	
  first	
  proposal	
  is	
  that	
  the	
  deficits	
  in	
  
rhythm	
  perception	
  found	
  by	
  Wieland	
  (2015)	
  were	
  due	
  to	
  the	
  difference	
  in	
  

	
  

27	
  

developmental	
  stage	
  of	
  the	
  populations	
  studied:	
  children	
  who	
  stutter	
  (CWS)	
  for	
  
Wieland	
  (2015),	
  in	
  contrast	
  to	
  adults	
  who	
  stutter	
  for	
  the	
  present	
  study.	
  Under	
  this	
  
explanation,	
  deficits	
  in	
  rhythm	
  perception	
  exhibited	
  in	
  CWS	
  may	
  disappear	
  by	
  
adulthood,	
  possibly	
  due	
  to	
  the	
  development	
  of	
  a	
  compensatory	
  strategy.	
  A	
  second	
  
possibility	
  for	
  the	
  contrasting	
  results	
  of	
  the	
  present	
  study	
  is	
  due	
  to	
  the	
  robust	
  
strength	
  of	
  the	
  manipulations	
  utilized	
  in	
  this	
  study.	
  There	
  is	
  a	
  possibility	
  that	
  
deficits	
  in	
  rhythm	
  and	
  timing	
  perception	
  in	
  PWS	
  might	
  be	
  subtle;	
  thus,	
  such	
  deficits	
  
might	
  not	
  have	
  been	
  observed	
  in	
  the	
  present	
  study	
  due	
  to	
  the	
  large	
  effect	
  sizes	
  
produced	
  by	
  the	
  manipulations	
  across	
  both	
  paradigms	
  used	
  in	
  this	
  study.	
  	
  The	
  third	
  
possible	
  reason	
  for	
  the	
  contradictory	
  results	
  is	
  that	
  this	
  study	
  utilized	
  perception	
  of	
  
rhythm	
  and	
  timing	
  in	
  speech,	
  while	
  Wieland	
  et	
  al.	
  (2015)	
  used	
  a	
  perception	
  task	
  
based	
  solely	
  on	
  tones,	
  without	
  speech.	
  It	
  is	
  possible	
  that	
  the	
  difference	
  in	
  tasks	
  
(tones	
  vs.	
  speech)	
  could	
  have	
  resulted	
  in	
  PWSs’	
  relying	
  on	
  a	
  different	
  set	
  of	
  
characteristics	
  for	
  speech	
  vs.	
  tones.	
  The	
  fourth	
  proposal	
  for	
  the	
  contradictory	
  
results	
  across	
  studies	
  could	
  be	
  that	
  the	
  manipulations	
  utilized	
  in	
  this	
  current	
  study	
  
did	
  not	
  rely	
  heavily	
  enough	
  on	
  the	
  internal	
  generation	
  of	
  rhythm	
  and	
  timing.	
  The	
  
generation	
  of	
  the	
  rhythms	
  and	
  timings	
  utilized	
  in	
  these	
  experiments	
  were	
  cued	
  
externally	
  through	
  the	
  prosodic	
  manipulations	
  of	
  distal	
  contexts	
  and	
  carried	
  
throughout	
  a	
  minimal	
  time-­‐span,	
  rather	
  than	
  fully	
  being	
  generated	
  internally.	
  	
  Thus,	
  
the	
  present	
  results	
  are	
  consistent	
  with	
  the	
  hypothesis	
  that	
  PWS	
  can	
  perceive	
  
rhythms	
  and	
  timings	
  cued	
  by	
  externally	
  provided	
  distal	
  prosodic	
  cues	
  but	
  that	
  they	
  
have	
  deficits	
  in	
  internal	
  rhythm	
  and	
  timing.	
  Future	
  research	
  should	
  address	
  this	
  and	
  

	
  

28	
  

further	
  investigate	
  with	
  a	
  paradigm	
  utilizing	
  purely	
  internally-­‐generated	
  rhythm	
  
and	
  timing	
  information.	
  
It	
  has	
  been	
  shown	
  that	
  internal	
  generation	
  of	
  rhythm	
  and	
  pacing	
  in	
  people	
  
who	
  stutter	
  (PWS)	
  can	
  be	
  influenced	
  by	
  perception	
  of	
  prosodic	
  aspects	
  of	
  their	
  own	
  
speech	
  (Hargrave,	
  Kalinowski,	
  Stuart,	
  Armson,	
  &	
  Jones,	
  1994;	
  Wingate,	
  1966).	
  
Guenther	
  (1994)	
  suggested	
  that	
  auditory-­‐motor	
  integration	
  is	
  used	
  for	
  spoken	
  
language	
  of	
  fluent	
  speakers	
  and	
  that	
  speech	
  production	
  occurs	
  by	
  integrating	
  
perceptions	
  of	
  self-­‐speech	
  into	
  ongoing	
  speech	
  production.	
  Previous	
  research	
  points	
  
to	
  a	
  problem	
  in	
  some	
  aspect	
  of	
  the	
  auditory-­‐integration	
  process	
  in	
  a	
  manner	
  that	
  
generates	
  disfluency.	
  There	
  is	
  evidence	
  that	
  shows	
  PWS	
  have	
  difficulty	
  with	
  this	
  
auditory-­‐motor	
  integration	
  of	
  speech	
  perceptions	
  into	
  speech	
  production	
  (Cai	
  et	
  al.,	
  
2014).	
  	
  These	
  findings	
  of	
  this	
  study	
  help	
  us	
  to	
  distinguish	
  among	
  three	
  possible	
  loci	
  
for	
  the	
  deficit	
  in	
  PWS	
  related	
  to	
  the	
  auditory-­‐motor	
  integration	
  of	
  prosody	
  in	
  speech	
  
identified	
  by	
  Cai	
  et	
  al.	
  (2014)	
  and	
  related	
  research.	
  One	
  is	
  the	
  possibility	
  of	
  difficulty	
  
in	
  the	
  perception	
  of	
  prosody.	
  A	
  second	
  possibility	
  is	
  difficulty	
  in	
  the	
  production	
  of	
  
prosodic	
  elements	
  in	
  speech.	
  A	
  third	
  possibility	
  is	
  difficulty	
  of	
  integrating	
  perception	
  
of	
  prosodic	
  elements	
  from	
  self-­‐speech	
  into	
  speech	
  productions.	
  These	
  possibilities	
  
are	
  discussed	
  below	
  with	
  reference	
  to	
  the	
  findings	
  of	
  the	
  present	
  study.	
  	
  
While	
  stuttering	
  is	
  thought	
  of	
  as	
  being	
  a	
  problem	
  in	
  the	
  production	
  of	
  speech,	
  
speech	
  perception	
  has	
  been	
  shown	
  to	
  play	
  an	
  important	
  role	
  in	
  the	
  speech	
  
production	
  of	
  PWS	
  (Hargrave,	
  Kalinowski,	
  Stuart,	
  Armson,	
  &	
  Jones,	
  1994;	
  Wingate,	
  
1966).	
  This	
  is	
  evidenced	
  by	
  the	
  effects	
  of	
  changing	
  the	
  auditory	
  feedback	
  of	
  pitch	
  or	
  
timing	
  on	
  the	
  fluency	
  of	
  PWS	
  with	
  delayed	
  auditory	
  feedback	
  (DAF)	
  devices	
  (Cai	
  et	
  

	
  

29	
  

al.,	
  2014).	
  The	
  possibility	
  of	
  a	
  problem	
  with	
  perception	
  of	
  prosodic	
  elements	
  of	
  
speech	
  provides	
  one	
  possible	
  explanation	
  for	
  the	
  underlying	
  deficits	
  in	
  PWS.	
  The	
  
results	
  of	
  the	
  present	
  study	
  suggest	
  that	
  the	
  perception	
  of	
  rhythm	
  and	
  timing	
  in	
  
adults	
  who	
  stutter	
  remains	
  intact.	
  	
  These	
  findings	
  suggest	
  that	
  the	
  problem	
  does	
  not	
  
seem	
  to	
  lie	
  within	
  the	
  perceptions	
  of	
  prosodic	
  elements	
  of	
  speech	
  in	
  PWS.	
  
The	
  hypothesis	
  of	
  difficulty	
  in	
  speech	
  production	
  is	
  another	
  potential	
  locus	
  of	
  
influence	
  of	
  prosody	
  in	
  underlying	
  deficits	
  of	
  PWS.	
  	
  If	
  the	
  problem	
  is	
  within	
  the	
  
motor	
  mechanisms	
  for	
  speech	
  production,	
  PWS	
  would	
  not	
  be	
  able	
  to	
  produce	
  
speech	
  that	
  exhibited	
  prosodic	
  elements	
  such	
  as	
  appropriate	
  rhythm,	
  timing,	
  or	
  
pitch.	
  PWS	
  can	
  however	
  demonstrate	
  appropriate	
  prosodic	
  elements	
  in	
  their	
  speech	
  
productions	
  with	
  the	
  assistance	
  of	
  externally	
  generated	
  rhythm	
  and	
  pacing	
  devices	
  
(Hargrave,	
  Kalinowski,	
  Stuart,	
  Armson,	
  &	
  Jones,	
  1994;	
  Wingate,	
  1966).	
  	
  The	
  fact	
  that	
  
PWS	
  can	
  exhibit	
  prosodic	
  elements	
  in	
  fluent	
  speech	
  with	
  assistive	
  fluency-­‐
enhancing	
  devices	
  leads	
  to	
  the	
  idea	
  that	
  the	
  problem	
  does	
  not	
  rely	
  in	
  solely	
  the	
  
production	
  of	
  these	
  prosodic	
  elements.	
  	
  Further	
  evidence	
  was	
  presented	
  in	
  a	
  recent	
  
poster	
  presentation	
  at	
  the	
  annual	
  conference	
  of	
  the	
  American	
  Speech-­‐Language-­‐
Hearing	
  Association	
  by	
  Wieland	
  et	
  al.	
  (2014)	
  that	
  showed	
  that	
  prosodic	
  aspects	
  of	
  
fluent	
  speech	
  in	
  children	
  who	
  stutter	
  were	
  similar	
  to	
  those	
  of	
  a	
  group	
  of	
  matched	
  
control	
  children.	
  These	
  results	
  suggest	
  that	
  an	
  underlying	
  deficit	
  in	
  production	
  of	
  
prosodic	
  elements	
  such	
  as	
  rhythm	
  and	
  timing	
  are	
  not	
  likely	
  to	
  be	
  a	
  core	
  deficit	
  in	
  
PWS.	
  
The	
  third	
  possibility	
  is	
  that	
  the	
  problem	
  lies	
  within	
  the	
  integration	
  of	
  
perception	
  into	
  speech	
  production.	
  This	
  possibility	
  is	
  supported	
  by	
  the	
  results	
  found	
  

	
  

30	
  

by	
  Cai	
  et	
  al.	
  (2014)	
  showing	
  a	
  deficit	
  in	
  PWS	
  in	
  the	
  integration	
  of	
  perturbed	
  auditory	
  
feedback	
  into	
  their	
  speech	
  productions.	
  	
  Cai	
  et	
  al.	
  (2014)	
  demonstrated	
  results	
  of	
  
weak	
  error	
  correction	
  of	
  speech	
  production	
  in	
  PWS	
  when	
  their	
  perceptions	
  of	
  
timing	
  had	
  been	
  disrupted.	
  	
  This	
  idea	
  of	
  difficulty	
  within	
  auditory-­‐integration	
  is	
  
supported	
  by	
  the	
  fluency-­‐inducing	
  effects	
  of	
  auditory-­‐feedback	
  devices	
  that	
  alter	
  the	
  
pitch	
  and	
  timing	
  in	
  auditory	
  feedback	
  of	
  self-­‐speech.	
  Without	
  altered	
  auditory-­‐
feedback,	
  the	
  incorporation	
  of	
  normal	
  feedback	
  into	
  speech	
  production	
  seems	
  to	
  be	
  
disrupted,	
  as	
  it	
  typically	
  results	
  in	
  disfluency.	
  Once	
  prosodic	
  elements	
  of	
  the	
  
auditory	
  feedback	
  is	
  manipulated,	
  fluency	
  is	
  often	
  enhanced,	
  supporting	
  the	
  idea	
  of	
  
a	
  disruption	
  in	
  the	
  integration	
  of	
  auditory-­‐feedback	
  (perception	
  of	
  self-­‐speech)	
  into	
  
speech	
  production.	
  
From	
  previous	
  literature	
  and	
  the	
  results	
  of	
  this	
  study,	
  the	
  perception	
  and	
  
production	
  of	
  prosodic	
  elements	
  of	
  speech,	
  each	
  considered	
  individually,	
  do	
  not	
  
display	
  a	
  problem	
  underlying	
  the	
  deficits	
  of	
  PWS.	
  	
  Thus,	
  the	
  present	
  work	
  supports	
  
the	
  idea	
  that	
  a	
  new,	
  fruitful	
  hypothesis	
  to	
  be	
  tested	
  involves	
  investigating	
  how	
  
prosodic	
  information	
  from	
  speech	
  perception	
  is	
  integrated	
  within	
  speech	
  
production.	
  While	
  this	
  study	
  primarily	
  focused	
  on	
  perception	
  of	
  prosodic	
  elements	
  
and	
  internal	
  rhythm	
  generation	
  in	
  PWS,	
  a	
  focus	
  of	
  research	
  on	
  the	
  auditory-­‐
integration	
  of	
  speech	
  perception	
  into	
  motor	
  production	
  of	
  speech	
  is	
  suggested	
  to	
  be	
  
most	
  productive	
  for	
  investigation	
  of	
  underlying	
  deficits	
  in	
  PWS	
  associated	
  with	
  
prosody.	
  	
  Despite	
  the	
  lack	
  of	
  understanding	
  of	
  overall	
  core	
  deficits	
  in	
  PWS,	
  the	
  
results	
  from	
  the	
  current	
  study	
  have	
  provided	
  evidence	
  of	
  an	
  intact	
  ability	
  of	
  PWS	
  to	
  
perceive	
  prosodic	
  elements	
  in	
  speech	
  and	
  to	
  internally	
  generate	
  rhythm	
  and	
  timing	
  

	
  

31	
  

following	
  distal	
  prosodic	
  cues	
  in	
  speech.	
  Research	
  defining	
  underlying	
  deficits	
  (or	
  
the	
  lack	
  thereof)	
  have	
  significant	
  clinical	
  implications,	
  since	
  core	
  strategies	
  or	
  
techniques	
  can	
  be	
  recognized	
  when	
  core	
  deficits	
  or	
  strengths	
  within	
  a	
  clinical	
  
population	
  are	
  recognized	
  and	
  understood.	
  .	
  The	
  results	
  of	
  this	
  study	
  shed	
  light	
  onto	
  
new	
  potential	
  areas	
  for	
  research	
  to	
  continue	
  investigation	
  of	
  an	
  underlying	
  deficit	
  
related	
  to	
  prosody	
  in	
  PWS.	
  	
  
	
  

	
  

	
  

32	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

APPENDICES	
  

33	
  

Appendix	
  A:	
  Experiment	
  1	
  Items	
  List	
  
	
  
1. 	
   banker	
  helpful	
  (tie	
  murder	
  bee/timer	
  derby)	
  	
  
2. 	
   kettle	
  heaven	
  (Tim	
  burrow	
  bow/timber	
  oboe)	
  	
  
3. 	
   pebble	
  dollar	
  (bar	
  lever	
  chew/barley	
  virtue)	
  
4. 	
   gossip	
  oyster	
  (pan	
  treaty	
  coy/pantry	
  decoy)	
  	
  
5. 	
   angry	
  index	
  (lay	
  birdie	
  fence/labor	
  defense)	
  	
  
6. 	
   chapter	
  elbow	
  (rue	
  beaver	
  gin/ruby	
  virgin)	
  	
  
7. 	
   kitchen	
  dealer	
  (may	
  beanie	
  grow/maybe	
  negro)	
  
8. 	
   hero	
  vacuum	
  (sell	
  early	
  gull/cellar	
  legal)	
  	
  
9. 	
   bullet	
  junior	
  (come	
  feeding	
  key/comfy	
  dinky)	
  	
  
10. 	
   liquid	
  perish	
  (broad	
  leasing	
  king/broadly	
  sinking)	
  	
  
11. 	
   lumpy	
  danger	
  (chair	
  eager	
  knee/cherry	
  gurney)	
  	
  
12. 	
   plasma	
  honey	
  (pigs	
  typo	
  low/pigsty	
  polo)	
  	
  
13. 	
   blanket	
  mounted	
  (ham	
  mercy	
  nick/hammer	
  scenic)	
  	
  
14. 	
   tourist	
  robin	
  (draw	
  musty	
  plea/drama	
  steeply)	
  	
  
15. 	
   trouble	
  wealthy	
  (limb	
  burner	
  sing/limber	
  nursing)	
  
16. 	
   nature	
  lazy	
  (faux	
  meaty	
  tour/foamy	
  detour)	
  	
  
17. 	
   lady	
  jacket	
  (bran	
  diesel	
  tree/brandy	
  sultry)	
  	
  
18. 	
   husband	
  lemon	
  (fan	
  seaman	
  cheese/fancy	
  munchies)	
  	
  
19. 	
   center	
  northern	
  (two	
  cancer	
  plus/toucan	
  surplus)	
  	
  
20. 	
   mixture	
  pleasure	
  (class	
  seedy	
  pose/classy	
  depots)	
  
	
  
	
  

	
  

34	
  

Appendix	
  B:	
  Experiment	
  2	
  Items	
  List	
  
	
  
1.	
  
The	
  murrays	
  are	
  (our)	
  favorite	
  
2.	
  
The	
  johnsons	
  are	
  (our)	
  rich	
  relatives	
  from	
  California	
  
3.	
  
The	
  accountants	
  are	
  (our)	
  wise	
  advisors	
  
4.	
  
We	
  won't	
  have	
  any	
  winter	
  (or)	
  wet	
  weather	
  
5.	
  
Marty	
  gave	
  him	
  a	
  dollar	
  (or)	
  twenty	
  last	
  week	
  
6.	
  
Fred	
  would	
  rather	
  have	
  (a)	
  summer	
  or	
  lake	
  
7.	
  
Mom	
  said	
  these	
  are	
  (our)	
  gray	
  gloves	
  
8.	
  
Sam	
  got	
  it	
  in	
  writing	
  before	
  (her)	
  last	
  wishes	
  
9.	
  
It's	
  not	
  long	
  before	
  (her)	
  rare	
  wit	
  
10.	
   John	
  didn’t	
  tell	
  the	
  junior	
  (or)	
  representative	
  about	
  it	
  
11.	
   The	
  snowsuits	
  are	
  (our)	
  only	
  
12.	
   The	
  Clines	
  are	
  (our)	
  good	
  friends	
  
13.	
   People	
  were	
  offended	
  after	
  (her)	
  rude	
  
14.	
   These	
  copy	
  machines	
  are	
  (our)	
  largest	
  
15.	
   Anne	
  wanted	
  to	
  see	
  (a)	
  very	
  funny	
  
16.	
   Susan	
  said	
  those	
  are	
  (our)	
  black	
  socks	
  
17.	
   New	
  laws	
  are	
  (our)	
  last	
  chances	
  for	
  a	
  change	
  in	
  climate	
  
18.	
   Deena	
  doesn't	
  have	
  any	
  leisure	
  (or)	
  time	
  
19.	
   Connor	
  knew	
  that	
  bread	
  and	
  butter	
  (are)	
  both	
  
20.	
   It's	
  not	
  easy	
  to	
  convey	
  (a)	
  likely	
  
21.	
   Sam	
  knew	
  there	
  (are)	
  apples	
  
22.	
   Todd	
  said	
  there	
  (are)	
  rooms	
  
23.	
   John	
  said	
  he	
  would	
  obey	
  (a)	
  rebel	
  
24.	
   Sally	
  might	
  try	
  (a)	
  liquid	
  
25.	
   The	
  sign	
  was	
  replaced	
  after	
  (her)	
  black	
  
26.	
   Don	
  must	
  see	
  the	
  harbor	
  (or)	
  boats	
  
27.	
   The	
  message	
  was	
  clear	
  after	
  (her)	
  blank	
  
28.	
   Tina	
  decorated	
  with	
  paper	
  (or)lace	
  in	
  most	
  rooms	
  
29.	
   Sue	
  said	
  there	
  (are)	
  lunches	
  
30.	
   Phil	
  and	
  Mary	
  are	
  (our)	
  young	
  cousins	
  
31.	
   Glenn	
  thought	
  his	
  friend	
  and	
  neighbor	
  (are)	
  like	
  plenty	
  
32.	
   Dave	
  asked	
  how	
  long	
  it	
  takes	
  to	
  repay	
  (a)	
  large	
  
33.	
   Rose	
  knew	
  that	
  there	
  (are)	
  lamps	
  
34.	
   Dawn	
  said	
  that	
  it's	
  easy	
  to	
  go	
  to	
  (a)	
  regular	
  
35.	
   Those	
  tickets	
  are	
  (our)	
  late	
  entries	
  
36.	
   John	
  called	
  her	
  sugar	
  (or)	
  honey	
  most	
  of	
  the	
  time	
  
37.	
   The	
  leaves	
  fell	
  after	
  (her)	
  green	
  
38.	
   They	
  were	
  sad	
  after	
  (her)	
  poor	
  
39.	
   The	
  Lees	
  are	
  (our)	
  old	
  friends	
  from	
  high	
  school	
  
40.	
   It's	
  not	
  long	
  before	
  (her)	
  bad	
  
	
  
	
  
	
  

	
  

35	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

36	
  

REFERENCES	
  	
  
	
  
Alm,	
  P.	
  A.	
  (2004).	
  Stuttering	
  and	
  the	
  basal	
  ganglia	
  circuits:	
  a	
  critical	
  review	
  of	
  
possible	
  relations.	
  Journal	
  of	
  communication	
  disorders,	
  37(4),	
  325-­‐369.	
  
	
  
Adams,	
  M.	
  R.,	
  &	
  Ramig,	
  P.	
  (1980).	
  Vocal	
  characteristics	
  of	
  normal	
  speakers	
  and	
  
stutterers	
  during	
  choral	
  reading.	
  Journal	
  of	
  Speech,	
  Language	
  and	
  Hearing	
  
Research,	
  23(2),	
  457-­‐469.	
  
	
  	
  
Bergmann,	
  G.	
  (1986).	
  Studies	
  in	
  stuttering	
  as	
  a	
  prosodic	
  disturbance.	
  Journal	
  of	
  
Speech,	
  Language	
  and	
  Hearing	
  Research,	
  29(3),	
  290-­‐300.	
  
	
  
	
  Besozzi,	
  T.	
  E.,	
  &	
  Adams,	
  M.	
  R.	
  (1969).	
  The	
  influence	
  of	
  prosody	
  on	
  stuttering	
  
adaptation.	
  Journal	
  of	
  Speech,	
  Language,	
  and	
  Hearing	
  Research,	
  12(4),	
  818-­‐
824.	
  
	
  	
  
Cai,	
  S.,	
  Beal,	
  D.	
  S.,	
  Ghosh,	
  S.	
  S.,	
  Guenther,	
  F.	
  H.,	
  &	
  Perkell,	
  J.	
  S.	
  (2014).	
  Impaired	
  timing	
  
adjustments	
  in	
  response	
  to	
  time-­‐varying	
  auditory	
  perturbation	
  during	
  
connected	
  speech	
  production	
  in	
  persons	
  who	
  stutter.	
  Brain	
  and	
  language,	
  
129,	
  24-­‐29.Chang,	
  S.-­‐E.,	
  &	
  Zhu,	
  D.	
  (2013).	
  Neural	
  network	
  connectivity	
  
differences	
  in	
  children	
  who	
  stutter.	
  Brain,	
  136(12),	
  3709-­‐3726.	
  	
  
	
  
Coull,	
  J.T.,	
  2004.	
  fMRI	
  studies	
  of	
  temporal	
  attention:	
  allocating	
  attention	
  within,	
  or	
  
	
  
towards,	
  time.	
  Cogn.	
  Brain	
  Res.	
  21,	
  216–226.	
  
	
  
Cutler,	
  A.,	
  Dahan,	
  D.,	
  &	
  van	
  Donselaar,	
  W.	
  (1997).	
  Prosody	
  in	
  the	
  comprehension	
  of	
  
spoken	
  language:	
  A	
  literature	
  review.	
  Language	
  and	
  Speech,	
  40,	
  141–201.	
  
	
  
Dilley,	
  L.	
  C.,	
  &	
  McAuley,	
  J.	
  D.	
  (2008).	
  Distal	
  prosodic	
  context	
  affects	
  word	
  
segmentation	
  and	
  lexical	
  processing.	
  Journal	
  of	
  Memory	
  and	
  Language,	
  59(3),	
  
294-­‐311.	
  
	
  
	
  Dilley,	
  L.	
  C.,	
  &	
  Pitt,	
  M.	
  A.	
  (2010).	
  Altering	
  context	
  speech	
  rate	
  can	
  cause	
  words	
  to	
  
appear	
  or	
  disappear.	
  Psychological	
  Science,	
  21(11),	
  1664-­‐1670.	
  	
  
	
  
Glover,	
  H.,	
  Kalinowski,	
  J.,	
  Rastatter,	
  M.,	
  &	
  Stuart,	
  A.	
  (1996).	
  Effect	
  of	
  instruction	
  to	
  
sing	
  on	
  stuttering	
  frequency	
  at	
  normal	
  and	
  fast	
  rates.	
  Perceptual	
  and	
  Motor	
  
Skills,	
  83,	
  511-­‐522.	
  
	
  
	
  Grahn,	
  J.A.,	
  &	
  Brett,	
  M.,	
  2007.	
  Rhythm	
  perception	
  in	
  motor	
  areas	
  of	
  the	
  brain.	
  J.	
  
Cogn.	
  Neurosci.	
  19,	
  893–906.	
  
	
  

	
  

37	
  

Grahn,	
  J.	
  A.,	
  &	
  McAuley,	
  J.	
  D.	
  (2009).	
  Neural	
  bases	
  of	
  individual	
  differences	
  in	
  beat	
  
perception.	
  Neuroimage,	
  47(4),	
  1894-­‐1903.	
  
doi:10.1016/j.neuroimage.2009.04.039	
  
	
  
Guenther,	
  F.	
  H.	
  (1994).	
  A	
  neural	
  network	
  model	
  of	
  speech	
  acquisition	
  and	
  motor	
  
equivalent	
  speech	
  production.	
  Biological	
  Cybernetics,	
  72(1),	
  43-­‐53.	
  
	
  
	
  Hall,	
  K.	
  D.,	
  Amir,	
  O.,	
  &	
  Yairi,	
  E.	
  (1999).	
  A	
  longitudinal	
  investigation	
  of	
  speaking	
  rate	
  
in	
  preschool	
  children	
  who	
  stutter.	
  Journal	
  of	
  Speech,	
  Language	
  and	
  Hearing	
  
Research,	
  42(6),	
  1367–1377.	
  	
  
	
  
Hargrave,	
  S.,	
  Kalinowski,	
  J.,	
  Stuart,	
  A.,	
  Armson,	
  J.,	
  &	
  Jones,	
  K.	
  (1994).	
  Effect	
  of	
  
frequency-­‐altered	
  feedback	
  on	
  stuttering	
  frequency	
  at	
  normal	
  and	
  fast	
  
speech	
  rates.	
  Journal	
  of	
  Speech,	
  Language,	
  and	
  Hearing	
  Research,	
  37(6),	
  1313-­‐
1319.	
  
	
  	
  
Harrington,	
  D.L.,	
  Haaland,	
  K.Y.,	
  1999.	
  Neural	
  underpinnings	
  of	
  temporal	
  processing:	
  
a	
  review	
  of	
  focal	
  lesion,	
  pharmacological,	
  and	
  functional	
  imaging	
  research.	
  
Rev.Neurosci.	
  10,	
  91–116.	
  
	
  
Harrington,	
  D.L.,	
  Haaland,	
  K.Y.,	
  Knight,	
  R.T.,	
  1998b.	
  Cortical	
  networks	
  underlying	
  
	
  
mechanisms	
  of	
  time	
  perception.	
  J.	
  Neurosci.	
  18,	
  1085–1095.	
  
	
  
Hazeltine,	
  E.,	
  Helmuth,	
  L.L.,	
  Ivry,	
  R.,	
  1997.	
  Neural	
  mechanisms	
  of	
  timing.	
  Trends	
  
Cogn.	
  Sci.	
  1,	
  163–169.	
  
	
  
Ingham,	
  R.	
  J.,	
  &	
  Carroll,	
  P.	
  J.	
  (1977).	
  Listener	
  judgment	
  of	
  differences	
  in	
  stutterers'	
  
nonstuttered	
  speech	
  during	
  chorus-­‐and	
  nonchorus-­‐reading	
  conditions.	
  
Journal	
  of	
  Speech,	
  Language	
  and	
  Hearing	
  Research,	
  20(2),	
  293-­‐302.	
  
	
  	
  
Kent,	
  R.	
  D.	
  (1984).	
  Stuttering	
  as	
  a	
  temporal	
  programming	
  disorder.	
  In	
  R.	
  F.	
  Curlee	
  &	
  
W.	
  H.	
  Perkins	
  (Eds.),	
  Nature	
  and	
  treatment	
  of	
  stuttering:	
  New	
  directions	
  (pp.	
  
283-­‐301).	
  San	
  Diego:	
  College-­‐Hill	
  Press.	
  
	
  
Morrill,	
  T.	
  H.,	
  Dilley,	
  L.	
  C.,	
  &	
  McAuley,	
  J.	
  D.	
  (2014).	
  Prosodic	
  patterning	
  in	
  distal	
  
speech	
  context:	
  Effects	
  of	
  list	
  intonation	
  and	
  f0	
  downtrend	
  on	
  perception	
  of	
  
proximal	
  prosodic	
  structure.	
  Journal	
  of	
  Phonetics,	
  46,	
  68-­‐85.	
  
	
  
Moulines,	
  E.,	
  &	
  Charpentier,	
  F.	
  (1990).	
  Pitch-­‐synchronous	
  waveform	
  processing	
  
techniques	
  for	
  text-­‐to-­‐speech	
  synthesis	
  using	
  diphones.	
  Speech	
  
Communication,	
  9(5-­‐6),	
  453–467.	
  
	
  
Nenadic,	
  I.,	
  Gaser,	
  C.,	
  Volz,	
  H.-­‐P.,	
  Rammsayer,	
  T.,	
  Hager,	
  F.,	
  Sauer,	
  H.,	
  2003.	
  Processing	
  
of	
  temporal	
  information	
  and	
  the	
  basal	
  ganglia:	
  new	
  evidence	
  from	
  fMRI.	
  Exp.	
  
Brain	
  Res.	
  148,	
  238–246.	
  
	
  
	
  

38	
  

Packman,	
  A.,	
  Code,	
  C.,	
  &	
  Onslow,	
  M.	
  (2007).	
  On	
  the	
  cause	
  of	
  stuttering:	
  Integrating	
  
theory	
  with	
  brain	
  and	
  behavioral	
  research.	
  Journal	
  of	
  Neurolinguistics,	
  20(5),	
  
353-­‐362.	
  doi:	
  10.1016/j.jneuroling.2006.11.001	
  
	
  
Packman,	
  A.,	
  Onslow,	
  M.,	
  Richard,	
  F.,	
  &	
  Van	
  Doorn,	
  J.	
  (1996).	
  Syllabic	
  stress	
  and	
  
variability:	
  A	
  model	
  of	
  stuttering.	
  Clinical	
  Linguistics	
  &	
  Phonetics,	
  10(3),	
  235-­‐
263.	
  
	
  	
  
Rao,	
  S.M.,	
  Mayer,	
  A.R.,	
  Harrington,	
  D.L.,	
  2001.	
  The	
  evolution	
  of	
  brain	
  activation	
  
during	
  temporal	
  processing.	
  Nat.	
  Neurosci.	
  4,	
  317–323.	
  
	
  
Salverda,	
  A.	
  P.,	
  Dahan,	
  D.,	
  Tanenhaus,	
  M.	
  K.,	
  Crosswhite,	
  K.,	
  Masharov,	
  M.,	
  &	
  
McDonough,	
  J.	
  (2007).	
  Effects	
  of	
  prosodically-­‐modulated	
  sub-­‐phonetic	
  
variations	
  on	
  lexical	
  competition.	
  Cognition,	
  105,466-­‐476.	
  
	
  
Schubotz,	
  R.I.,	
  Friederici,	
  A.D.,	
  von	
  Cramon,	
  D.Y.,	
  2000.	
  Time	
  perception	
  andmotor	
  
timing:	
  a	
  common	
  cortical	
  and	
  subcortical	
  basis	
  revealed	
  by	
  fMRI.	
  
Neuroimage	
  11,	
  1–12.	
  
	
  
Toyomura,	
  A.,	
  et	
  al.,	
  Effect	
  of	
  an	
  8-­‐week	
  practice	
  of	
  externally	
  triggered	
  speech	
  on	
  
basal	
  ganglia	
  activity	
  of	
  stuttering	
  and	
  fluent	
  speakers,	
  NeuroImage	
  (2015),	
  
http://dx.doi.org/10.1016/j.neuroimage.2015.01.024	
  
	
  
Wieland,	
  E.	
  A.,	
  Dilley,	
  L.	
  C.,	
  Burnham,	
  E.	
  B.,	
  Redford,	
  M.,	
  &	
  Chang,	
  S-­‐E.	
  (2014).	
  
Prosodic	
  characteristics	
  of	
  fluent	
  speech	
  by	
  children	
  who	
  do	
  and	
  do	
  not	
  
stutter.	
  Poster	
  presented	
  at	
  the	
  Annual	
  Meeting	
  of	
  the	
  American	
  Speech-­‐
Language	
  Hearing	
  Association,	
  Orlando,	
  FL,	
  November	
  20-­‐22,	
  2014	
  
	
  
	
  Wieland,	
  E.	
  A.,	
  McAuley,	
  J.	
  D.,	
  Dilley,	
  L.	
  C.,	
  &	
  Chang,	
  S.	
  E.	
  (2015).	
  Evidence	
  for	
  a	
  
rhythm	
  perception	
  deficit	
  in	
  children	
  who	
  stutter.	
  Brain	
  and	
  language,	
  144,	
  
26-­‐34.Wingate,	
  M.	
  E.	
  (1966).	
  Prosody	
  in	
  stuttering	
  adaptation.	
  Journal	
  of	
  
Speech,	
  Language,	
  and	
  Hearing	
  Research,	
  9(4),	
  550-­‐556.	
  	
  
	
  
Wingate,	
  M.	
  E.,	
  &	
  Howell,	
  P.	
  (2002).	
  Foundations	
  of	
  stuttering.	
  J	
  Acoust	
  Soc	
  Am(112),	
  
1229-­‐1241.	
  
	
  	
  
World-­‐Health-­‐Organization.	
  (2010).	
  International	
  statistical	
  classification	
  of	
  
diseases	
  and	
  related	
  health	
  problems.	
  	
  	
  Retrieved	
  4/30/13,	
  from	
  
http://apps.who.int/classifications/icd10/browse/2010/en#/F98.5	
  
	
  

	
  

39