Phonological Encoding in Cluttering
Emily O.
Charley F.
1
Adams ,
Allen A.
1
Montgomery ,
Kenneth O. St.
2
Louis ,
& Dirk-Bart den
1
Ouden
of Communication Science and Disorders, University of South Carolina, Columbia, SC, USA, 2Department of Speech Pathology and Audiology, West Virginia University, Morgantown, WV, USA
[email protected]
Auditory Monitoring
Auditory Monitoring
Methods
Participants
• Participants (see Table 1) completed the Self-Awareness of Speech Index
(St. Louis & Adkins, 2005) and Predictive Cluttering Inventory (Daly, 2006)
• Control participants had no speech, language, or hearing deficits
• One participant, diagnosed with both cluttering and stuttering, has been
included in graphs for comparison, but was excluded from statistical analyses
Mean Age
Handedness (R/L/B)
M/F
SASI
PCI
Clutterers (n=4)
39.0
2/1/1
1/3
2.87
105.0
Stutterers (n=3)
40.7
2/1/0
2/1
2.71
71.0
Stutterer/Clutterer (n=1)
45.0
1/0/0
1/0
2.56
78.0
Controls (n=6)
35.5
5/1/0
3/3
2.86
31.2
Table 1. Participant demographics.
• Participants pressed the spacebar as soon as a tone was presented, with a
random interstimulus interval of either 200ms, 500ms, 1000ms, or 1500ms
• Allowed us to rule out any basic motor response differences between groups
Results
Naming and PALPA Tests
• Low error rates on naming of experimental
target words, across groups (Mean %
error: clut. = 3.6, stut. = 1.2, clut/stut. = 0,
cont. = 0.6) (Figure 1)
• No group differences were found for
repetition of real words (PALPA 9), word
rhyme judgment (PALPA 15), or final
segmentation (PALPA 17)
• All 3 experimental groups made more
errors than controls on initial
segmentation (PALPA 16), rhyme
judgment requiring picture selection
(PALPA 14), and naming (PALPA 53)
• Clutterers made most errors on repetition
of nonwords (PALPA 9)
Phoneme Monitoring
• Tasks were a partial replication of Sasisekaran et al., (2006)
• Participants monitored for the presence or absence of a target phoneme
(consonant + schwa, e.g. /bə/) during silent picture naming
• Participants were familiarized with the words and required to name them
correctly prior to the experiment
• 28 bisyllabic words were used, with the target phoneme occurring in one of
four positions, C1VC2C3VC4 (e.g. “p1ig2l3et4”)
• Phonemes to monitor: /p/, /t/, /k/, /b/, /d/, /g/, /m/, /n/, /s/, /ʃ/, /r/, /l/, /f/, /v/,
which were balanced among position within the words
20
Clutterers
Stutterers
Controls
Clut/Stut
10
5
0
Figure 1. Mean errors for each group on naming of target
words and PALPA subtests.
Reaction Time (ms)
40
30
Clutterers
Stutterers
Controls
Stut/Clut
20
10
0
2
3
4
Position
Figure 2. Mean RT – Phoneme Monitoring
Mean
Clutterers
Stutterers
Controls
Stut/Clut
20
10
2
3
4
Mean
0
1
2
3
4
Mean
Position
Figure 4. Mean RT – Auditory Monitoring
Figure 5. Mean % error – Auditory Monitoring
Simple Motor Task
• No differences were found for any group in basic motor response
abilities (Means: clut. = 252ms, stut. = 248ms, clut/stut. = 309ms,
cont. = 251ms)
Discussion and Conclusions
50
1
30
Position
15
• For RT, the main effect of Group was not significant as determined by a Kruskal
Wallis test (χ2 = 4.96, p = 0.084) (Clut. M = 1219.53ms, Stut. M = 1390.37, Cont.
M = 1070.8; Figure 2). Despite nonsignificant group results, pairwise comparisons
were made, showing that only stutterers’ RTs were significantly longer than those
of the controls (χ2 = 4.27, p = .0389)
• For % error, the main effect of Group was not significant (Clut. M = 20.33%, Stut.
M = 19.94%, Cont. M = 8.04%; χ2= 3.49, p = 0.175; Figure 3). Still, pairwise
comparisons were made and stutterers exhibited a nonsignificant trend to make
more errors than controls (χ2= 2.86, p = 0.0906)
• Clutterers and stutterers both had mean error rates of approximately 20%
1800
1600
1400
1200
1000
800
600
400
200
0
40
1
Phoneme Monitoring
Pre-tests
• Participants first completed subtests from the Psycholinguistics Assessments
of Language Processing in Aphasia (PALPA, 1992)
• Auditory Repetition of Words and Nonwords (9), Rhyme Judgment
Requiring Picture Selection (14), Word Rhyme Judgment (15),
Phonological Segmentation of Initial Sounds (16), Phonological
Segmentation of Final Sounds (17), and Picture Naming (53)
25
50
1800
1600
1400
1200
1000
800
600
400
200
0
% Error
Simple Motor Task
• For RT, the main effect of Group was not significant (Clut. M = 580.76ms,
Stut. M = 729.21ms, Cont. M = 769.08ms; Figure 4) (χ2 = 2.40, p = 0.30
indicating that the mean RT of each group was similar.
• For %errors, there was a significant main effect of Group (χ2 = 6.3824, p =
0.0411). Follow up tests revealed that stutterers (M = 15.36%) made more
errors than controls (M = 4.28%) (χ2 = 3.85, p =0.0499), and that clutterers
(M = 14.13%) also made more errors than controls (χ2= 4.67, p = 0.0309)
• Clutterers and stutterers again were very similar in their error rates
Reaction Time (ms)
• Whether or not cluttering is a speech fluency disorder with a language
component or is a separate language disorder remains debated
• Cluttering may be a motor speech disorder especially considering some of
the major characteristics such as rate deviations, excessive coarticulation,
and abnormal pausing (Ward, 2011)
• It has been suggested that speech errors in cluttering arise because of lack
of planning or formulation time, possibly related to phonological encoding
(Van Zaalen, Wijnen, & Dejonckere, 2009)
• Stutterers have been shown to be slower on a phoneme monitoring task,
suggesting pre-articulatory problems at the level of phonological encoding
(Sasisekaran et al., 2006)
• Purpose of study: to determine whether clutterers’ phonological
manipulation abilities are different than that of controls and of
stutterers
• If clutterers’ speech output pattern reflects their phonological
encoding abilities, they may be faster and more error-prone on a task
that taps into this level of processing, viz. self-monitoring for speech
sounds (phonemes) in inner/covert speech
• However, if the internal monitoring step is absent, or impaired, in
clutterers, there is no reason to expect faster times when they are
forced to deliberately monitor the internal speech plan in an
experimental task
• An auditory monitoring task was developed to closely approximate the phoneme
monitoring task
• Participants monitored for one of four tones (500Hz, 1000Hz, 2000Hz, or 2500Hz)
among a sequence of four tones (same frequencies with addition of 1500Hz as a
possibility) using a similar procedure as in the phoneme detection task
• Allowed general auditory monitoring skills to be assessed, which does not involve
any phonological or lexical processing
% Error
Introduction
% Error
1Department
1
Garnett ,
1
2
3
4
Mean
Position
Figure 3. Mean % error – Phoneme Monitoring
References:
Daly, D. A. (2006). Predictive Cluttering Inventory.
Kay, J., Lesser, R., & Coltheart, M. (1992). Psycholinguistic assessment of language processing in Aphasia. London: Psychology Press.
Postma, A., & Kolk, H. (1993). The covert repair hypothesis: Prearticulatory repair processes in normal and stuttered disfluencies. Journal of
Speech and Hearing Research, 36(3), 472–487.
Sasisekaran, J., De Nil, L. F., Smyth, R., & Johnson, C. (2006). Phonological Encoding in the Silent Speech of Persons Who Stutter. Journal
of Fluency Disorders, 31, 1-21.
St. Louis, K. O. & Atkins, C. P. (2005). Self-awareness of Speech Index. Morgantown, WV: Populore.
Van Zaalen, Y., Wijnen, F., & Dejonckere, P. (2009). Language planning disturbances in children who clutter or have learning disabilities.
International Journal of Speech and Language Pathology, 11, 496-508.
Ward, D. (2011). Motor speech control and cluttering. In D. Ward & K. Scaler Scott (Eds.), Cluttering: A handbook of research, intervention, and
education. Hove, UK: Psychology Press.
• The present study replicated the findings of Sasisekaran et al. (2006)
by demonstrating that stutterers have slower reaction times than
controls during phoneme monitoring, but not auditory monitoring or a
simple motor task.
• Contrary to Sasisekaran et al. (2006) the stutterers in this study did
make significantly more errors than the controls during both tasks
• Clutterers made numerically more errors than controls during the
phoneme and complex auditory monitoring task – nearly the same
number as stutterers
• Under the assumption that the present study is low on power, we
therefore suggest phonological encoding is disrupted in clutterers,
leading to errors in the speech plan
• To that extent, the errors in cluttering may be generated during
phonological encoding, just like stuttering errors as accounted for by
the Covert Repair Hypothesis (CRH, Postma & Kolk, 1993)
• In contrast to what the CRH says about stuttering disfluencies being
side effects of internal error repairs (hypermonitoring), we propose
that in clutterers, internal monitoring of the prearticulatory speech plan
may be disrupted (hypomonitoring), in addition to the phonological
encoding deficit
• These two deficits in the clutterer’s system allow the errors to
advance to the final phonetic output, while the absence of timeconsuming internal monitoring accounts for the increased speech
rate in cluttering
• While not of primary interest in this study, clutterers’ relatively poor
performance on the nonword repetition task in the PALPA suggests
further research in this area is warranted
• PALPA results for all 3 experimental groups for segmentation and
rhyme judgment may also warrant further research on these tasks
• It is important to interpret the findings of this study with some caution
due to low participant numbers
• This ongoing study aims to increase power by continuing recruitment