P-center perception in children with developmental dyslexia: do low level auditory deficits underlie reading, spelling, and language impairments?

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
P-CENTER PERCEPTION IN CHILDREN WITH DEVELOPMENTAL DYSLEXIA:
DO LOW LEVEL AUDITORY DEFICITS UNDERLIE READING, SPELLING, AND
LANGUAGE IMPAIRMENTS
by
Rachel Lynette Beattie
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF ARTS
(PSYCHOLOGY)
December 2008
Copyright 2008 Rachel Lynette Beattie
ii
Table of Contents
List of Tables
Abstract
Introduction
Methods
Results
Discussion
Future Directions
Conclusions
References
iii
iv
1
11
21
29
32
34
35
iii
List of Tables
Table 1: Means and Standard Deviations on 21
the Qualification Measures for Each
Group
Table 2: Pearson’s Correlations Between 22
Subtests in the Dinosaur Task
Table 3: Pearson’s Correlations Between the 23
Measures of the Test Battery
Table 4: Means and Standard Deviations on 25
the Dinosaur Task for Each Group
Table 5: 95% Confidence Intervals for the 27
Dinosaur Tasks Based on the
Percentile Bootstrap Method with
20% Trimmed Means
iv
Abstract
A recent theory proposed by Goswami et al (2002), dubbed the P-Center
hypothesis, posits that the characteristic cognitive and behavioral patterns observed in
developmental dyslexia are a result of low-level auditory impairments. Previous studies
have found that children with developmental dyslexia perform worse on P-center
perception tasks when compared to chronological age matched controls, younger reading
level controls have intermediate thresholds, and this deficit has also been observed in
children with specific language impairment. The current study found similar trends in
children aged between 7-16, in that children with dyslexia and dyslexia with additional
language difficulties had less sensitive mean thresholds on the P-center perception tasks
than chronological age matched controls. If these trends persist, then more ecologically
valid stimuli and more sophisticated methods should be used to explore the P-center
deficit in children with developmental dyslexia.
1
In general, individuals with developmental dyslexia can be described as having
difficulties with reading and spelling not directly attributable to low intelligence, lack of
adequate reading instruction, or obvious sensory or neurological damage (Goswami,
Thomson, Richardson, Stainthorp, Hughes, Rosen, and Scott, 2002; Snowling, 2000,
Ramus, Rosen, Dakin, Day, Castellote, White, and Frith, 2003). Along with difficulties
in reading and spelling, characteristic impairments seen in dyslexic readers also include
impaired phonological awareness (Morris et al, 1998), rapid automatized naming (Wolf
and Bowers, 1999), and phonological short term memory (Chiappe, Hasher, and Siegel,
2000). These phonological processing impairments are observed for the majority of
dyslexic individuals, which supports the most prominent cognitive theory of
developmental dyslexia: the phonological core deficit hypothesis (Stanovich and Siegel,
1994; Chiappe, Stringer, Siegel, and Stanovich, 2002). This theory posits that dyslexic
adults and children have difficulty perceiving, creating, maintaining, and retrieving
phonological representations, which in turn affects other cognitive areas associated with
reading, such as phonological awareness and rapid automatized naming skills.
Development of these phonological representations is a fundamental step in literacy
acquisition, but if these representations are impaired, as they are in dyslexia, reading and
spelling development will be delayed (Gallagher, Frith, and Snowling, 2000).
However, the underlying mechanisms that impair the perception, creation,
maintenance, and retrieval of phonological representations are still debated (Ramus et al,
2003). A new theory proposed by Goswami et al (2002) explains that the characteristic
cognitive and behavioral patterns observed in dyslexia are a result of low-level auditory
2
impairments in perceiving amplitude envelope onset cues, particularly P-centers.
According to Scott (1998), P-centers can be defined as perceptual moments of occurrence
characterized by rapid changes in amplitude in the onset of both speech and non-speech
sounds. This property of sound is said to play a role in both the perception and production
of regularity in auditory stimuli. Accordingly, Goswami et al’s P-center theory claims
that deficits in processing these cues would result in impaired phonological
representations. The main purpose of this study was not only to replicate Goswami et al’s
findings that link P-center perception to reading and spelling development; but also to
extend our methodology to address theoretical gaps left by this theory and other
perspectives of dyslexia that have implicated an underlying auditory impairment.
The P-center hypothesis is not the first theory to purport that auditory
impairments play a crucial role in the etiology of developmental dyslexia (Tallal, 1980;
Stein, 2001). Specifically, these previous perspectives have claimed that dyslexic
individuals have difficulty perceiving and integrating rapid auditory stimuli. Making use
of tasks involving rapid frequency detection and temporal order judgment, these studies
typically find that only a small subset of dyslexic individuals in any given study, 25-30%
at the most, exhibit such impairments (Ramus et al, 2003). Despite some support for
these theories (see Halliday and Bishop, 2005), most attempts to replicate these rapid
auditory perception difficulties have failed (Chiappe, Stringer, Siegel, and Stanovich,
2002; Bailey and Snowling, 2002, Ramus et al, 2003; Gibson, Hogben, & Fletcher,
2006).
3
On the other hand, more favorable findings have been obtained for a newer
generation of experiments that have used a variety of amplitude modulations (Witton et
al, 2002), onset variations (Menell, McNally, and Stein, 1999), or other temporal
variations of spectral energy that did not solely focus on rapid auditory modulations
(Lorenzi, Dumont, and Fullgrabe, 2000). These studies have found that both children and
adults with dyslexia demonstrate less sensitivity to these acoustic properties and have
higher variability in task performance when contrasted with controls. Moreover,
perception of these cues has also been linked to reading and spelling ability (Menell,
McNally, and Stein, 1999). In an event-related potential study of 6-month-old infants,
Leppanen et al (2002) found that infants with a high familial risk of dyslexia had
different responses when perceiving consonant duration changes relative to those children
without a family history of dyslexia. When combined with their other findings that at-risk
infants also differ in their responses to changes in vowel duration and consonant-vowel
syllables (Leppanen et al, 1999; Guttorm et al, 2001), these results suggest that at-risk
infants may have difficulty representing durational aspects of the temporal envelope,
which are crucial in perceiving P-centers (Vos, Mates and van Kruysbergen, 1995).
Although dyslexic individuals appear to have more difficulty perceiving and
representing fine-grained aspects of the sound envelope, referred to as the temporal
envelope, relative to normal readers, the relationship between these deficits and the
impoverished phonological representations characteristic of dyslexia is not fully
understood. One of the main goals of the P-center hypothesis is to link this non-speech
4
specific mechanism to the developmental processes underlying language, reading and
spelling development.
In the speech stream, P-centers are associated with the production and perception
of the onsets of vowels, which have been implicated in the extraction of speech rhythm
(Goswami, 2003). Rapid changes in the temporal envelope, such as those inherent in
sound onsets, are considered to be more important for perception than other properties of
the envelope (Heil, 2003). Thus, the ability to detect P-centers in speech appears to play a
crucial role in extracting the rhythm from speech.
The extraction of speech rhythm is an invaluable tool for reading development as
it allows the speech stream to be segmented into syllables, words, and phrases. This
sensitivity to rhythm is part of a phonological subsystem, called prosody, which also
encompasses the tempo and stress of a language (Whalley and Hansen, 2006). Prosodic
cues are crucial for children's acquisition of spoken language since these cues provide
structure and highlight the different meanings contained in the speech stream. Sensitivity
to prosodic cues also plays a prominent role in reading development. Whalley and
Hansen (2006) found that 8 to 9 year old children's prosodic skills accounted for unique
variation in word reading ability and reading comprehension after controlling for
phonological awareness and general rhythmic sensitivity. The ability to extract rhythm,
tempo, and stress from speech was not simply due to general awareness of rhythm or
phonological awareness, but required phonological processing as well. Moreover, many
of the words in standardized reading and spelling tests are multi-syllabic. Since lexical
stress is not overtly indicated in the printed English language, children who have
5
difficulties extracting speech rhythm may be unable to fluently join together syllables
within multi-syllabic words and thus, may show more severe impairments when reading
longer words (Wade-Woolley and Wood, 2006)
Additionally, deficits in speech rhythm sensitivity have been linked to difficulty
in spelling. In two experiments conducted by Wood (2006), 5, 6, and 7-year-old children
were presented with familiar words, some of which were mispronounced. When the
metrical speech patterns of familiar words were reversed, an alteration that directly
manipulates speech rhythm, the children had more difficulty identifying the
mispronounced words than with other alteration types. Furthermore, there appeared to be
a developmental continuum with the 7 year olds scoring higher than the 5 year old
children on mispronounced word identification. In two separate analyses, the metrical
stress sensitivity scores accounted for unique variance in spelling scores after
phonological awareness or vocabulary scores had been taken into account. Moreover,
Wood found that poor rhyme awareness, another impairment typically seen in dyslexics
(Snowling, 1998), was associated with poor metrical sensitivity after accounting for age.
In addition to speech rhythm sensitivity, the development of appropriate stress
assignment patterns is also considered to be an aspect of supra-segmental phonology that
is important for literacy development (Whalley and Hansen, 2006). The same
phonological segment of a word is represented differently when it is stressed versus when
it is unstressed and the ability to use differences in lexical stress to parse the speech
stream is developed as early as 7 months (Curtin, Mintz, and Christiansen, 2005). If word
stress perception is impaired, then phonological representations will be affected
6
accordingly. De Bree, Wijnen, and Zonneveld (2006) investigated word stress acquisition
in Dutch at-risk dyslexic children relative to children without a family history of dyslexia.
The task required three-year-old children to repeat non-words with four types of word
stress patterns: regular, irregular, highly irregular, and prohibited. The regular and
irregular patterns were considered to be valid in Dutch whereas the prohibited type was
an invalid pattern not naturally found in the language. Compared to the control group, the
at-risk children changed the word and syllable structure significantly more often in order
to create a simpler and more regular word stress. The control children scored highly for
both the regular and irregular stress patterns whereas the at-risk children only had high
repetition scores on the regular stress words. Since children develop the regular stress
patterns earlier than irregular ones, the at-risk children’s difficulties with the irregular
patterns indicates a delay in word stress acquisition.
P-centers are closely linked to the perception and production of rhythm in the
auditory stream, thus a deficit in perceiving these auditory cues could cause
desensitization to prosodic information, ultimately leading to impaired phonological
representations. Although P-centers are linked to the onsets of vowels in speech, they
provide a non-speech specific way to measure the perception of rhythm in the auditory
stream (Goswami et al, 2002). Accordingly, the auditory perception tasks created to
measure P-center perception do not necessarily have to be speech based and can be used
cross-linguistically or before language abilities are fully developed. In order to detect the
location of a P-center, the listener uses relative, event-based cues rather than employing a
global approach (Scott, 1998). By altering the onset of the amplitude envelope, referred
7
to as rise time, and the duration of a sound in an auditory task, experimenters can shift the
location of the P-center in the sound (Vos, Mates and van Kruysbergen, 1995). For
example, shorter rise times and longer durations cause the P-center to be shifted towards
the offset. Moreover, variation of only the rise times can affect the perception of distinct,
discrete ‘beats’ in the auditory stream. Slow, long rise times lead to the perception of a
continuous sound that varies in loudness whereas the use of shortened, quick rise times
prompts the perception of a continuous sound with a loud beat occurring at the same rate
as the modulation.
Goswami et al (2002) used non-speech tasks that manipulated rise time of
amplitude modulated sequences to measure P-center perception in four groups of
children: dyslexic readers, reading age matched controls, chronological age-matched
controls, and precocious readers. After accounting for age, IQ, and vocabulary, the P-
center perception tasks accounted for 25% of independent variance in reading and
spelling. Once phonological skills were also accounted for, the rise time sensitivity
measure remained the only significant predictor of the auditory task and explained 9% of
unique variance in reading and spelling. Moreover, there appeared to be a developmental
continuum for the beat perception task with dyslexic readers performing significantly
worse than chronological age-matched controls and the precocious readers performing
significantly better than these controls. Despite the interesting developmental
implications, this finding has yet to be replicated.
All of the experiments that have used similar non-speech P-center perception tasks
have essentially replicated Goswami et al’s (2002) initial findings that dyslexic adults and
8
children are less sensitive to amplitude modulations relative to controls (Richardson et al,
2004;Thomson et al, 2006; Pasquini et al, 2007; Thomson and Goswami, 2008). This
pattern of deficits has also been observed in cross-linguistic studies indicating that a P-
center perception deficit is present in dyslexic individuals despite lexical variations
(French: Muneaux et al 2004; Finnish: Hamalianen et al, 2005; Hungarian: Csépe et al,
2006). However, although these studies have found decreased sensitivity to P-centers in
individuals with dyslexia relative to normal reading controls, performance has been
variable on the measures designed to assess P-center perception (Ahissar and Onganian,
2008). For example, both Pasquini et al (2008) and Thomson et al (2006) found that the
best predictor of individual differences in reading and spelling was the 1-ramp rise time
task, a P-center perception subtest where the initial rise time of the sound is manipulated.
However, Richardson et al (2004) found that the 2-ramp rise time task, a measure with
two repetitions of the same P-center stimuli, was a better predictor of unique variance in
the children’s reading and spelling scores than the 1-ramp task. Before any conclusions
can be made about which P-center perception task best predicts unique variation in
reading and spelling, more studies need to be conducted that use these measures.
In addition to the between-groups analyses, Richardson et al (2004) also performed
a within-group analysis for the dyslexic children. Following the example of Ramus et al
(2003), children who achieved a score below the 5
th
percentile for their age were
considered to be deficient in P-center perception. This analysis revealed that 15 of the 24
(or 62.5%) dyslexic subjects were below the 5
th
percentile on the 1-ramp rise time task
9
and 10 of those 15 individuals (or 67%) were also below the 5
th
percentile on the other
rise time task.
Moreover, Corriveau et al (2007) obtained similar within-group results for
children with Specific Language Impairment (SLI). SLI is a developmental disorder
characterized by difficulties in expressive and receptive oral language despite normal
non-verbal intelligence and no evidence of sensory or neurological impairments (Stark
and Tallal, 1981). Similar to developmental dyslexia, studies have failed to find auditory
deficits consistent with previous sensory theories in all children with SLI (Bishop et al,
1999; McArthur and Bishop, 2001) and the exact role of these possible auditory
impairments in the development of language problems is still being debated (Rosen,
2003). When Corriveau et al compared the performance of chronological age matched
controls and SLI children, 71.4% of the SLI group fell below the 5
th
percentile on the 1-
ramp rise time task. Although none of the SLI children fell below the 5
th
percentile on the
2-ramp ramp task, 52% fell below the 25
th
percentile and 76% were below the 50
th
percentile.
Despite finding the same P-center impairment, the children in Corriveau et al’s
study had no developmental history of dyslexia and the dyslexic subjects recruited in the
other P-center perception studies had no specific speech and language deficits (Goswami
et al, 2002; Richardson et al, 2004; Thomson et al, 2006). While it is valuable to consider
these two populations separately, developmental dyslexia is often comorbid with SLI
(Catts et al, 2005) and thus, it would be ecologically valid to include a group of dyslexic
individuals with additional oral language comprehension and production difficulties
10
(referred to as the D-LD group). Moreover, sensory studies that have included a D-LD
group have found that these participants exhibit the most severe impairments (Ahissar et
al, 2006; Sperling et al, 2005) and therefore, the P-center deficit may be exacerbated if
SLI and dyslexia are comorbid. However, a D-LD group has not been included in any of
the P-center studies.
In response to the findings of the P-center hypothesis, this study was designed not
only to replicate the presence of a P-center deficit in children diagnosed with
developmental dyslexia, but also to compare several different groups on the rise time
manipulated measures. In addition to the standard dyslexic, or reading disabled (RD)
group and chronological age matched (CA) group, we also included a reading level (RL)
matched group. The RL group, which is typically three years younger than the D and CA
groups and matched to the D group on reading, was included in Goswami et al’s (2002)
study, but has not been included in further experiments. By comparing the dyslexic
individuals to younger normally achieving readers, this will reveal whether the P-center
perception difficulty is a by-product of poor reading, or potentially a source of poor
reading. Furthermore, a D-LD group, matched to the D group on reading but impaired on
measures of oral language comprehension and production, was also included.
Comparisons between the D group and the D-LD group would clarify whether the P-
center deficit is characteristic of specific reading disability or more generally of children
with reading and language delays.
11
Methods
Participants
We studied twenty-four participants, 11 normal readers (CA group, mean age = 11
years - 7 months, 9 boys, 2 girls), 7 dyslexic readers (D group, mean age = 12 years – 6
months, 7 boys, 2 girls), 3 dyslexic readers with additional oral language difficulties (D-
LD group, mean age = 12 years – 4 months, 3 girls), and 3 reading level controls (RL
group, mean age = 7 years – 10 months, 1 boy, 2 girls). Over a period of thirty months,
these children were recruited via Craigslist, flyers in university staff mailboxes on the
USC campus, and teacher nominations based on reading ability. Before scheduling a
testing session, a telephone screening session was done with the parent or guardian of the
child. The purpose of this screening was to exclude possible subjects that had a history of
serious emotional or behavioral problems, uncorrected auditory or vision impairments,
neurological problems, or were not fluent English speakers. Additionally, the child’s
parent or guardian and primary teacher completed the Child Behavior Checklist (CBCL;
Achenbach, 1991) and Disruptive Behavior Rating Scale (Barkley and Murphy, 1998).
These scales were administered in order to screen out any child who had untreated
behavioral or emotional disorders.
After assenting to the experiment, the child was tested on a series of qualifying
tasks (see the descriptions of these tasks and group requirements in the measures section).
If the participant’s scores did not qualify them for membership in any of the subject
groups, the child was excluded from further testing. All children who qualified for the
study completed the full test battery listed under the measures section.
12
Measures: Qualification Measures and Composite Estimates
Word Identification
This subtest of the Woodcock-Johnson Tests of Achievement (WJ-III) required the
child to read words aloud (Woodcock, McGrew, and Mather, 2001). All subtests of the
WJ-III were normed with 4,873 children between kindergarten and 12
th
grade. The word
identification measure had a split test reliability between 0.88 and 0.97 for ages 7 to 16.
For our study, a percentile score of 40% or above on this test is required for membership
in the CA group. Both the D and D-LD groups were characterized by achieving a
percentile score of 25% or lower on this subtest or on the Word Attack subtest. Those in
the RL group were matched to the D and D-LD children on their raw scores from Word
Identification.
Word Attack
Word Attack is another subtest of the WJ-III in which the child is asked to read
nonsense words, out loud. This subtest had split test reliability coefficients between 0.78
and 0.92 for the ages in our sample. A percentile score of 40% or above on this test was
required for membership in the CA group and both the D and D-LD groups were
characterized by achieving a percentile score of 25% or lower on this subtest or on the
Word Identification subtest.
Non-Verbal IQ Estimate
To estimate non-verbal IQ, two subtests of the Weschler Intelligence Scale for
Children: Third Edition (WISC-III; Weschler, 1991) were used: Block Design and
Picture Completion. In the former, the child was asked to use a set of blocks to replicate
13
designs either constructed by the experimenter or printed in a booklet. The designs
became increasingly complex and performance on this subtest was timed. The block
design subtest has an internal consistency reliability between 0.77 and 0.92 for children
aged between 7 and 16, with an average of 0.87. Picture Completion required the
participant to tell or point to the most important part that was missing from a picture
within 20 seconds. This subtest from the WISC-III was also used to estimate non-verbal
IQ and had an internal consistency reliability coefficients that ranged between 0.72 and
0.84 for ages 7 to 16, with a mean of 0.77.
The standard scores on these two subtests were averaged to provide the non-verbal
IQ estimate in this study. To qualify for any group in this study, the child needed to
obtain an average of 7 or higher.
Verbal IQ Estimate
The verbal IQ estimate was calculated from performance on the Vocabulary and
Similarities subtests of the WISC-III. In the Vocabulary subtest, the experimenter
prompted the child to define words orally. For ages 7 to 16, the internal consistency
reliability coefficients for the Vocabulary subtest ranged between 0.79 and 0.91, with a
mean of 0.87. In Similarities, the tester instructed the child to explain orally how two
words were similar. The internal consistency reliability coefficients for ages 7 to 16
ranged between 0.74 and 0.84, with a mean of 0.81.
Like the non-verbal IQ estimate, the verbal IQ estimate was calculated by
averaging the standard scores achieved on Vocabulary and Similarities. To be in the D,
CA, and RL groups, the children had to obtain an average standard score of at least 7.
14
Oral Language Composite
In response to Corriveau et al’s (2007) findings with SLI children, we included
three measures of oral language production and comprehension to assess whether the P-
center deficit was also observed in those children who have both reading and oral
language difficulties, the D-LD group. To qualify for the D-LD group, the children had to
score below the 30
th
percentile on two of the three tasks in addition to meeting the other
criteria for membership in the D group. Moreover, an oral language composite was
calculated in order to facilitate regression and correlation analyses. To calculate the oral
language composite, we averaged the standard scores obtained on three measures of oral
language production and comprehension: Concepts and Directions, Recalling Sentences,
and Receptive One-Word Picture Vocabulary Test (ROWPVT).
The Concepts and Directions subtest from the CELF-3, Clinical Evaluation of
Language Fundamentals – Third Edition (Semel, Wiig, and Secord, 1995), asked children
to listen carefully to a set of instructions about a set of geometric shapes before being
asked to perform the actions in the order in which they were said. For this subtest, the
internal reliability coefficients ranged between 0.59 to 0.86 for ages 7 to 16, with a
median of 0.76. Recalling Sentences, also from the CELF-3, required the subjects to
listen to a sentence and then to repeat it exactly as it was said. The internal reliability
coefficient for ages 7 to 16 ranged between 0.83 to 0.91, with a median of 0.90. The final
measure used for the language estimate was ROWPVT (Brownell, 2000). In this
receptive oral language task, children were shown sets of four pictures and told to
indicate in each case the picture that best matched the word said by the experimenter. For
15
our sample, the internal consistency reliability coefficient alphas ranged from 0.95 to
0.98 with a median of 0.97.
Measures: Reading and Spelling Tasks
Fluency
For this measure of reading fluency from the WJ-III, the subject was given three
minutes to silently read a series of sentences and for each one, the child was asked to
indicate if the sentence was accurate or nonsensical by circling Y or N, respectively, next
to each sentence. For this timed test, split-test reliability coefficients were between 0.87
and 0.94 for ages 7 to 16.
Spelling
In this WJ-III subtest, the subject was prompted to spell words. For the ages in
our sample, the split-test reliability coefficients ranged from 0.87 to 0.92.
Exception words
In this measure devised by Bailey et al (2004), the child was asked to read a series
of 70 words aloud that got progressively lower in printed word frequency. All words
contained one or more violations of typical spelling-sound correspondences (e.g., said,
beauty, anchor, silhouette). The split-half reliability with a Spearman-Brown correction
for length was 0.94.
Test of Word Reading Efficiency (TOWRE) - Sight Word Efficiency
This TOWRE subtest consisted of two lists of words that the participant had to
read out loud as quickly, accurately, and fluently as possible. (Torgesen, Wagner, and
16
Rashotte, 1999). Since this measure was a speeded test, Cronbach’s alpha was not used
and instead reliability was calculated using alternative-form coefficients, in which the
subjects took both forms of the test in one session and the scores on Form A were
correlated to the scores on Form B. The coefficients for the Sight Word Efficiency
subtest ranged between 0.88 and 0.97 and the coefficients for the total word reading
efficiency score ranged between 0.91 and 0.97 for ages 7 to 16.
TOWRE - Phonological Decoding
Unlike the Sight Word Efficiency subtest, the Phonological Decoding subtest
required the participants to read through two lists of non-words as quickly, fluently, and
accurately as possible. Again, alternate-form coefficients were calculated to determine
the reliability of this subtest of the TOWRE. For the ages in our study, the coefficients
ranged between 0.91 and 0.97 for this subtest.
Gray Silent Reading Test (GSRT)
For this task, children silently read short stories and answered comprehension
questions about them. (Wiederholt and Blalock, 2000). The coefficient alphas for the
ages in our sample ranged from 0.93 to 0.96, with a median of 0.94.
Gray Oral Reading Test (GORT)
This task required children to read stories out loud as fluently, quickly, and
accurately as possible (Wiederholt and Bryant, 1992). The internal consistency reliability
coefficients for children at ages 7 to 16 ranged between 0.92 and 0.93 for the rate score,
0.86 and 0.92 for the accuracy score, and between 0.90 and 0.94 for the passage score.
17
Measures: Phonological Processing Tasks
Elision
In this Comprehensive Test of Phonological Processing (CTOPP) subtest, the
child was asked to say a word and then asked to repeat the word with either a given
syllable or phoneme missing. (Wagner, Torgesen, and Rashotte, 1999). The internal
consistency reliability coefficients ranged between 0.81 to 0.92 for ages 7 to 16, with a
median of 0.89.
Segmenting Non-words
For this CTOPP subtest, participants heard a nonsense word and were asked to
repeat the word and then to repeat the phonemes of the word one sound at a time.
(Wagner, Torgesen, and Rashotte, 1999). The Cronbach’s alphas for the ages in our study
ranged from 0.87 to 0.91, with a median of 0.89.
Non-word Repetition
This CTOPP subtest asked subjects to listen to nonsense words and then to repeat
each nonsense word exactly as it was said on the tape (Wagner, Torgesen, and Rashotte,
1999). In this subtest, the internal consistency reliability coefficients were between 0.73
and 0.80, with a median of 0.785.
Measures: Tests of Other Cognitive Abilities
Calculations
In this subtest of the WJ-III, children were asked to solve a series of math
problems. The split-test reliability coefficients for ages 7 to 16 ranged between 0.80 and
0.87.
18
Memory for Digits
In this short-term memory subtest of the CTOPP, subjects heard a series of
numbers, ranging from two to nine in a set, and were asked to repeat the digits in the
order that they heard them (Wagner, Torgesen, and Rashotte, 1999). For this subtest, the
internal consistency reliability coefficients were between 0.72 and 0.81 for the ages in
our study, with a median of 0.78.
Measures: Auditory Tasks
The Dinosaur task, designed by Dorothy Bishop at Oxford University, was used to
assess the children’s P-center perception. A computer located in a sound attenuated booth
presented the pairs of stimuli through a pair of noise canceling headphones and the
loudness of the stimuli in each task was set at 73 dB, apart from the comparison tones
presented in the Intensity Discrimination subtest. Each part of this program used an
adaptive two interval forced choice (2IFC) format and consisted of a maximum of 40
trials in which the sound pairs were separated by an interval of 500ms. The participants
received continuous feedback during the trials. The threshold value for each task was
determined by finding the point where the child responded correctly 75% of the time over
the last four reversals. This was accomplished by using a more virulent form of
Parameter Settings by Sequential Estimation (PEST; Findlay, 1978).
The Intensity discrimination Task:
This task, based on the perception tasks used by Ivry and Keele (1989), measured
the child’s ability to distinguish between tones that differed in loudness. The pairs of
19
sound in this paradigm were 50ms, 1 kHz pure tones. A standard tone of 73dB was
always presented alongside a second sound that was adaptively selected from a set of 31
pure tones. The loudness of the second tone ranged between 73 and 81.1 dB, with 0.27
dB between each step. The child was asked to indicate which dinosaur was associated
with the louder tone. The threshold scores obtained from this task referred to the
difference, in decibels, needed between two tones to identify the louder tone with 75%
accuracy.
The 1-ramp rise time task:
This experiment assessed the participant’s ability to perceive P-centers using a
simple manipulation of rise time. Children were presented with pairs of 500 Hz pure
tones, amplitude-modulated at 0.7 Hz with a depth of 50%. Each sound had a total
duration of 800 ms and a fixed linear fall time of 50 ms. The standard tone consisted of a
300 ms linear rise time envelope and a 450 ms steady state. The comparison tone,
adaptively selected from a set of 40 stimuli, had a linear rise time envelope that increased
logarithmically from 15 to 300 ms and a steady state that was varied accordingly so that
the total duration always equaled 800 ms. The child’s task was to indicate which dinosaur
had the steepest rise time (“Which Dinosaur was louder at the beginning?”). For this rise
time measure, threshold values referred to the 40 comparison tone levels. Accordingly,
those with lower thresholds were sensitive to smaller differences between the comparison
and standard tones compared to those with threshold scores closer to 40.
20
Figure 1: Part a represents two conditions in the 1-ramp rise time task, the 15ms and
300ms rise times, respectively. Part b represents the 2-ramp rise time task with 300ms
and 15ms rise times, respectively. (Taken from Richardson et al, 2004)
The 2-ramp rise time task:
This paradigm assessed the participant’s P-center detection abilities on a more
complex level. Like the 1-ramp task, the comparison tone was adaptively selected from a
set of 40 stimuli with a rise time that was logarithmically varied from 15 to 300 ms
whereas the standard sound had a linear rise time of 300 ms. However, each tone was
cycled 2.5 times, 3570 ms long, with a fixed linear fall time of 350 ms. The child was
asked to indicate which of the two stimuli had the sharpest rise-times (“Which dinosaur
made the sound with the sharper beat?”). Like the other rise time measure, the threshold
scores were out of 40, in reference to the number of comparison tone stimuli.
21
Results
Descriptive statistics
The mean standard scores on the qualifying measures and composite estimates for
each group were calculated using SPSS 16.0 Graduate Student Version and are listed in
Table 1. The main purpose of including the RL group is to investigate whether the
performance on the experimental measures is due to maturational factors, reading ability,
or a combination of both. However, the current participants recruited for that group have
both mean Word ID and Word Attack scores that exceed the means of all the other
groups, including the CA group. Thus, the RL group will be excluded from the current
group analyses and membership in that group will be reevaluated and accurately matched
once all the participants needed for the D and D-LD groups have been recruited.
Table 1: Means and Standard Deviations on the Qualification Measures for Each Group
Group Word ID Word Attack Verbal IQ
estimate
Non-Verbal
IQ estimate
Language
Estimate
CA
(n = 11)
106.77
(6.94)
118.18
(9.79)
12.86
(2.24)
12.55
(3.27)
45.75
(6.10)
D
(n = 7)
87.00
(13.10)
85.29
(10.80)
10.57
(3.19)
10.21
(3.77)
40.38
(4.50)
D-LD
(n= 3)
80.00
(1.00)
93.67
(2.31)
9.33
(2.08)
9.00
(0.87)
33.89
(1.26)
RL
(n=3)
115.67
(3.22)
120.33
(10.41)
10.83
(1.26)
9.83
(2.02)
44.67
(4.41)
22
Correlation Analyses
A series of simple bi-variate correlations were conducted in order to explore the
potential links between the Dinosaur subtests and the other measures in our test battery.
As seen in Table 2, both rise time tasks were moderately correlated (p = 0.003) with each
other. Additionally, the 1-ramp rise time measure was significantly correlated with
intensity discrimination (p = 0.042). Since the intensity discrimination task is thought to
be a measure of attentional capacity (Richardson et al, 2004), it may be that the type of
attention needed to the stimuli in the intensity discrimination task is also needed for the
tones in the 1-ramp rise time task, but not the 2-ramp rise time measure. This possibility
will be explored further using the regression analyses outlined later in this section.
Table 2: Pearson’s Correlations Between Subtests in the Dinosaur Task
Intensity
discrimination
1-ramp rise
time
2-ramp rise
time
Intensity discrimination 1 0.428* 0.233
1-ramp rise time 0.428* 1 0.597**
2-ramp rise time 0.233 0.597** 1
P ≤ 0.05 (2-tailed) = *
P ≤ 0.01 (2-tailed)= **
To investigate potential relationships between the Dinosaur task and the other
measures in our battery, we ran a series of simple Pearson’s correlations, which are
displayed below in Table 3.
23
Table 3: Pearson’s Correlations Between the Measures of the Test Battery
Intensity Discrimination 1-ramp rise time 2-ramp rise time
Word ID .252 .041 -.333
Word Attack .273 .083 -.391
TOWRE Sight
Word Efficiency
.194 -.283 -.442
TOWRE
Phonological
Decoding
.163 -.265 -.378
Spelling .283 -.150 -.332
Calculations .424 -.240 -.397
Fluency .184 -.263 -.365
Segmenting Non-
words
.338 .185 .020
Non-word
Repetition
.389 .109 -.072
Memory for
Digits
.402 -.195 -.327
Phoneme Elision .388 .044 -.230
Exception Words
(raw score)
.092 -.215 -.410*
Verbal IQ
Estimate
.019 -.094 -.252
Non-verbal IQ
Estimate
.002 -.142 -.612**
Language
Estimate
.056 -.018 -.216
GSRT .130 -.051 -.224
GORT – Rate
Score
-.047 .215 .064
GORT –
Accuracy Score
-.413 -.334 .484
P ≤ 0.05 (2-tailed) = *
P ≤ 0.01 (2-tailed)= **
Based on the trends of these preliminary bi-variate correlations, the 2-ramp rise
time task was negatively related to performance on word and non-word reading,
exception word reading (p = 0.47), spelling, fluency, language, verbal IQ, working
24
memory, and non-verbal IQ (p = 0.001). There were no strong trends observed between
the 2-ramp rise time task and all measures of phonological processing, though phoneme
elision appeared to have a slightly inverse relation to the 2-ramp rise time task. While the
1-ramp rise time task did show some negative links to the same measures as the 2-ramp
measure, the trends are less compelling for the 1-ramp rise time. Moreover, most of the
other measures in the battery did not exhibit strong links to intensity discrimination,
except for calculations and memory for digits. However, no definitive conclusions should
be made based upon these preliminary correlations because Pearson’s r requires a larger
sample size to be robust enough to detect all the links within our test battery.
Group Differences
The means and standard deviations for the rise time and intensity discrimination
tasks are listed per group in Table 4. The threshold values for the rise time tasks were
calculated based on the forty levels of stimuli in the task. If the values were closer to
forty, that indicated lower sensitivity to rise time manipulations than those scoring closer
to zero. Based on these means, both the D and D-LD groups appear to have lower P-
center thresholds than the CA controls. The D and D-LD groups appear to differ in their
thresholds on the 1-ramp and 2-ramp rise measures; however, the pattern of deficits may
change once the sample size is larger. For the intensity discrimination thresholds, these
values refer to the difference between two tones, measured in decibels, that the child
needed to accurately identify the loudest tone 75% of the time. Values closer to twelve
25
indicated that the child was less sensitive to differences in intensity than a subject scoring
closer to zero. The intensity discrimination thresholds are roughly similar between the
groups and for all groups, there was a high amount of variability on performance in both
the rise-time tasks.
Table 4: Means and Standard Deviations on the Dinosaur Task for Each Group
Group 1-ramp rise time 2-ramp rise time Intensity
Discrimination
CA (n = 10) 15.81 (10.80) 17.66 (11.60) 3.92 (3.22)
D (n = 7) 16.85 (11.36) 26.81 (8.99) 2.69 (1.68)
D-LD (n = 3) 20.77 (14.22) 24.60 (14.33) 2.47 (0.64)
Due to the small sample size, the D and D-LD groups were pooled to increase
power. To compare the performance of the pooled D and D-LD group to that of the CA
group, we used the percentile bootstrap method with 20% trimmed means. This robust
method is very practical when sample sizes are small and there is high variability in the
data (Wilcox, 2003). The percentile bootstrap method samples, with replacement, from
the collected data set. This process is repeated at least a thousand times (β) and the
sample statistics, trimmed means in this instance, are calculated from these generated
bootstrap samples (X
t1
*, .... , X
tß
*). Since the percentile bootstrap method does not work
well with sample means, we used trimmed means instead. Trimmed means differ from
typical sample means in that it removes a certain percentage, as defined during the
26
analysis, of the data from the top and bottom of the distribution (i.e. the lowest 10% and
highest 10%). When using trimmed means with the percentile bootstrap method, at least
20% trimming is recommended in order to have accurate control over Type I error, even
with samples as small as 11 (Wilcox, 2003).
To test the significance of a hypothesis, the bootstrap trimmed means need to be
placed in ascending order [(X
t(1)
≤... ≤X
t(β) )
] before calculating a 95% confidence interval
based on these bootstrap statistics: (X
t
*
(L+1)
, X
t
*
(µ)
), where L = α(ß)/2 and µ = ß - L. The
null hypothesis, H
0
: µ
t
= µ
0
, is rejected if this confidence interval does not contain the
hypothesized value µ
0
. In the case of group comparisons, if the 95% confidence interval
for one group does not overlap with the 95% confidence interval obtained for the other
group, then the null hypothesis can be rejected.
The statistical program r was used to perform this percentile bootstrap method with
20% trimmed means to compare performance on the Dinosaur subtests between the
pooled D and D-LD group and the CA group. Based on our initial data sample, we cannot
conclude that there were significant group differences on any of the Dinosaur task
subtests. As seen below in Table 5, the 95% confidence intervals for the pooled D and D-
LD group and the CA group overlap for every subtest. Once the full sample is recruited,
group differences will be reevaluated and extended to include all four groups.
27
Table 5: 95% Confidence Intervals for the Dinosaur Tasks Based on the Percentile
Bootstrap Method with 20% Trimmed Means
Intensity Discrimination 1-ramp rise time 2-ramp rise time
Group
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Lower
Bound
Upper
Bound
Pooled D and D-
LD
1.61 3.63 8.68 27.42 18.86 33.41
CA 2.46 6.19 10.19 23.98 12.50 26.90
Regression Analyses
When the sample size is large enough to support regression analyses with adequate
Type I error control, a series of regression analyses will be conducted. These regression
analyses were designed to determine if performance on the P-center perception tasks
accounts for unique variance in the characteristic impairments observed in developmental
dyslexia even after other possible mechanisms (i.e. short-term memory, phonological
processing) have been taken into account. The dependent variables for all of the proposed
analyses will consist of a phonological processing component (a composite of the CTOPP
measures), the verbal IQ estimate, the language estimate, a word reading and spelling
composite variable (composed of Word ID, spelling, GORT, GSRT, fluency, TOWRE
sight word efficiency, and exception words), a non-word reading variable (consisting of
Word Attack and TOWRE phonological decoding efficiency), and calculations.
28
In the first regression analysis, 1) age and 2) non-verbal IQ will be entered before
the three dinosaur subtests. For the second set of regression analyses, we will take into
account 1) age, 2) non-verbal IQ, and 3) intensity discrimination before entering either
the 1-ramp or 2-ramp rise time measures. The intensity discrimination task, which did not
show group difference trends, has been used as an attentional measure in previous
regression analyses done in the P-center literature since this task has similar instructions,
stimuli, and trial length to the rise time measures (Richardson et al, 2004). If performance
on the rise time measures were due to attentional impairments, these tasks should not
account for any additional variance after intensity discrimination is entered.
The third set of regression analyses will take into account 1) age, 2) non-verbal IQ,
and 3) working memory before either rise time task. The P-center perception tasks
require working memory, a characteristic deficit in developmental dyslexia (Chiappe,
Hasher, and Siegel, 2000), in order to retain a representation of the first tone in a set
while attending to the second stimulus. If the P-center perception tasks are simply tapping
into working memory capacity rather than an impaired low level auditory impairment,
then any unique variance previously accounted for by the rise time measures would
overlap with that accounted for by working memory. Additionally the fourth series of
regression analyses will take into 1) age, 2) non-verbal IQ, and 3) phonological
processing before entering either of the rise time measures. If the 1-ramp or 2-ramp rise
time tasks still account for additional variance in the literacy measures after phonological
processing is entered, then it would mean that P-center perception contributes to reading
and spelling above and beyond its connection to phonological processing abilities.
29
Within Group analysis
Once the anticipated sample size for each group is achieved, a within-group
analysis of the D and D-LD groups will be performed. This analysis will be based on
Ramus et al (2003) and Richardson et al’s (2004) concept of deviance, in which subjects’
performance on the subtests of the Dinosaur task will be compared against that deemed
typical for their age. Any subject falling below the 5th percentile of the CA group
performance on a given task will be classified as deviant for that measure. By using this
procedure, we can better classify individuals with developmental dyslexia into subsets
based on those who show deviance in one, two, or none of the P-center detection tasks.
Discussion
This study aimed not only to replicate previous findings in the P-center literature,
but also to explore theoretical gaps about the developmental trajectory and the
relationship of this deficit to both reading and language difficulties. In general, our initial
results indicate that children with dyslexia appear to have less sensitive P-center
perception thresholds than chronologically aged matched controls, which replicates
Goswami et al’s (2002) findings. Like Richardson et al (2004), the 2-ramp rise time
measure was linked to most of the reading, spelling and language measures; however, no
definitive conclusions can be made about the predictive value of any measure. The small
sample size and subsequent matching issues with the younger readers prevented
exploration of the developmental trajectory of P-center perception ability. Although the
30
pooled D and D-LD groups had less sensitive thresholds on the P-center perception tasks
than the CA controls; we could not investigate whether this deficit was exacerbated by
the presence of additional language difficulties due to the small sample size.
Differential findings on the dyslexic group’s performance on the rise time tasks
have been reported before in the P-center perception literature, with some studies
reporting more severe deficits on one measure and not another (Thomson et al, 2006).
While our current findings support Richardson et al’s (2004) results with children
diagnosed with developmental dyslexia, Corriveau et al (2007) found that SLI children’s
thresholds on the 1-ramp rise time measure were the highest. Once the full sample size is
achieved and the regression analyses are conducted, it will be interesting to see if the 2-
ramp rise time measure accounts for the most unique variance in reading, spelling, and
language or if the 1-ramp rise time measure appears to be a better predictor of individual
differences. Also, it may be that one rise time measure is more predictive of individual
differences in one domain, i.e. oral language, whereas the other rise time task more
reliably accounts for unique variance in another cognitive are, i.e. reading. Although the
rise time subtests both assess P-center perception, they are inherently different on certain
dimensions. For example, the 1-ramp rise time measure contains shorter stimuli than the
2-ramp rise time task and thus, successful performance on the latter measure might be
more dependent on the subject’s attentional capacities than the 1-ramp rise time task
because the trials are longer. Such differences may prove more problematic for one
population over another and full group comparisons will help clarify the pattern of
deficits on these subtests.
31
Additionally, once we have recruited enough participants, a series of regression
analyses will be run that take into account age, non-verbal IQ, working memory,
attention, and phonological processing before considering the unique variance that the
rise time measures account for. These analyses are conducted to ensure that performance
on the rise time tasks does not overlap with the variance accounted for by other abilities.
If Goswami et al’s (2003) hypothesis is right, then the thresholds achieved on the P-
center tasks should account for additional variance in the reading and spelling measures.
Also, if the rise time measures account for unique variance in the language estimate, this
would replicate Corriveau et al’s (2007) results and provide support for the role of P-
center perception deficits in the development of oral language difficulties.
Moreover, the within groups analysis will clarify whether the lower sensitivity on
these tasks were due to a small subset of children with P-center impairments or were a
broader characteristic of all children with developmental dyslexia in this sample. When
replicating previous auditory hypotheses (i.e. Tallal ,1980; Stein, 2001), typically only
25-30% of the dyslexic group exhibits impairments on any given auditory task (Ramus et
al, 2003). If the prospective within-group analysis replicates those done by Richardson et
al (2004) and Corriveau et al (2007), then we can expect that roughly 66% of the dyslexic
children in this sample would exhibit impairments on the rise time measures. The high
correlation between the rise time tasks and the reading, spelling, and language measures
in this study lend support to Goswami et al.’s (cite a couple of papers here) claims that a
low level auditory deficit may underlie the characteristic deficits seen in dyslexia.
Moreover, although the links between the P-center perception tasks and the phonological
processing measures were not significant in the present analysis, there was a trend
32
associating poorer phoneme elision with less sensitivity to rise time manipulations. If
future analyses find a significant inverse relation between the P-center tasks and
phonological processing, this would support Goswami et al’s (2002) notion that
perception of P-centers contributes to the perception, creation, maintenance, and retrieval
of phonological representations, which is crucial to reading and spelling development.
Thus, this theory would not be at odds with the most prominent cognitive theory of
dyslexia and the majority of studies that find phonological awareness and processing
impairments in most individuals with developmental dyslexia.
Future Directions
Since P-centers are associated with the production and perception of the onsets of
vowels, which have been implicated in the extraction of speech rhythm (Goswami, 2003),
this theory provides another pathway from prosodic perception to reading disability that
deserves further exploration. Prosody, the tempo, stress, and rhythm of a language, is
linked to reading and spelling abilities (Whalley and Hansen, 2006) and is especially
important when reading multi-syllabic words in English since lexical stress is not overtly
indicated in the orthography (Wade-Woolley and Wood, 2006). P-center perception and
multi-syllabic word reading have not been directly compared in any empirical study.
Multi-syllabic words contain multiple vowel onsets, and P-centers, and thus may provide
a more ecologically valid way to explore P-center perception than the current rise time
tasks. Moreover, music provides another ecologically valid way of studying dyslexic
individuals’ sensitivity to manipulations in rise time and duration. Previous studies have
found that individuals with dyslexia show impaired perception of music rhythm and note
33
reading (Overy, 2003; Ganschow, Lloyd-Jones, and Miles, 1994). Also, Forgeard et al
(2008) found that melodic and rhythmic musical discrimination abilities accounted for
unique variance in phonemic awareness. Including both ecologically valid speech and
non-speech measures of P-center perception in the same test battery could better clarify
the role of P-center perception in reading development, especially if the study sampled
both children and adults with a wide range of reading skills.
As early as 7 months, infants are able to use prosodic cues to parse spoken language
(Curtin, Mintz, and Christiansen, 2005). However, these skills are impaired in 3 year old
children with a high familial history of developmental dyslexia (De Bree, Wijnen, and
Zonneveld, 2006) and prosodic and P-center perception may be less sensitive starting
even earlier in development. However, the question of how these perception
discrepancies initially arise remains unanswered. One possible explanation is that
children with developmental dyslexia have deficient neuronal responses to amplitude
envelope onset cues. Joris, Schreiner, and Rees (2003) posit that temporal modulations,
such as those in the amplitude envelope onsets, play a critical role in detecting,
discriminating, identifying, parsing, and localizing acoustic signals. Moreover, Heil and
Neubauer (2001, 2003, 2004) argue that when integrating an auditory signal over time,
the listener uses the sound’s pressure envelope and thus, amplitude modulations are
important when representing acoustic stimuli.
If a neuronal mechanism underlies the reduced sensitivity that dyslexic children
have towards P-centers and other prosodic cues, then this deficit might be able to be
localized in the brain using modern imaging techniques. Based on both single-cell
recording with animals and magneto encephalography (MEG) studies with adults, there
are neurons that selectively respond to different points of amplitude onsets (Biermann
and Heil (2000). It could be that these neurons fire differently when dyslexic listeners
34
perceive amplitude onsets than when normal reading controls attend to the same stimuli.
In addition to MEG, functional magnetic resonance imaging (fMRI) techniques have also
been used in an attempt to localize where amplitude onsets are processed in the brain
(Giraud et al, 2000). For example, Hall et al. (2000) found activation in the planum
temporale caudal to primary auditory cortex when subjects attended to amplitude
modulations. Not only do these imaging experiments help strengthen the theoretical
underpinnings of the P-center hypothesis, they also provide new methodology to further
explore its claims. Both MEG and fMRI could be used to investigate the differences in
the biological correlates of P-center perception in individuals with dyslexia as compared
to normal reading controls.
Conclusions
The trends observed in our initial analyses replicate the previous findings that
children with developmental dyslexia have less sensitive P-center perception thresholds
than chronologically aged matched controls. Also, in addition to being associated with
reading, spelling, and language development, initial analyses indicate that performance
on the P-center perception tasks was linked to phonological processing, which supports
Goswami et al’s (2002) claims. Once data collection is complete, we will be able to
explore the developmental trajectory of P-center perception and whether this low level
auditory cue is characteristic of broader language difficulties. If the current trends
persist, then further exploration of this auditory mechanism should be explored with more
ecologically valid speech and non-speech stimuli and also with methodology that can
better assess the nature of this impairment at the neuronal level.
35
References
Achenbach, T.M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991
Profile. Burlington, VT: University of Vermont Department of Psychiatry.
Ahissar, M., Lubin, Y., Putter-Katz, H., and Banai, K. (2006). Dyslexia and the Failure to
Form a Perceptual Anchor. Nature Neuroscience, 9, 1558-1564.
Ahissar, M. and Onganian, Y. (2008). Response to Ziegler: The Anchor is in the Details.
Trends in Cognitive Science, 12 (7), 245-246.
Bailey, P.J. and Snowling, M.J. (2002). Auditory Processing and the Development of
Language and Literacy. British Medical Bulletin, 63, 135-146.
Bailey, C. E., Manis, F. R., Seidenberg, M.S., Peterson, W., and Seidenberg, M.S.
(2004). Variation Among Developmental Dyslexics: Evidence from a Printed-Word-
Learning Task. Journal of Experimental Child Psychology, 87, 125-154.
Barkley, R. A. and Murphy, K. R. (1998). Attention-Deficit Hyperactivity Disorder. A
Clinical Workbook (ed 2). New York: Guilford Press.
Biermann, S. and Heil, P. (2000). Parallels Between Timing of Onset Responses in Single
Neurons in Cat and Evoked Magnetic Fields in Human Auditory Cortex. The Journal of
Neurophysiology, 84 (5), 2426-2439.
Bishop, D. V. M., Carlyon, R. P., Deeks, J. M., and Bishop, S. J. (1999). Auditory
Temporal Processing Impairment: Neither Necessary Nor Sufficient for Causing
Language Impairment in Children. Journal of Speech, Language, and Hearing Research,
42, 1295–1310.
Brownell, R. (ed) (2000). Receptive One-Word Picture Vocabulary Test. Novato, CA:
Academic Therapy Publications.
Catts, H. W., Adlof, S. M., Hogan, T. P., and Weismer, S. E. (2005). Are Specific
Language Impairment and Dyslexia Distinct Disorders? Journal of Speech, Language,
and Hearing Research, 48, 1378–1396.
Chiappe, P., Hasher, L., and Siegel, L.S., (2000). Working Memory, Inhibitory Control,
and Reading Disability. Memory and Cognition, 28(1), 8-17.
Chiappe, P., Stringer, R., Siegel, L.S., and Stanovich, K.E. (2002). Why the Timing
Deficit Hypothesis Does Not Explain Reading Disability in Adults. Reading and Writing:
An Interdisciplinary Journal, 15, 73-107.
36
Corriveau et al, 2007 Basic Auditory Processing Skills and Specific Language
Impairment: A New Look at an Old Hypothesis. 50 (3), 647 –666.
Curtin, S., Mintz, T. H., and Christiansen, M. H. (2005). Stress Changes the
Representational Landscape: Evidence from Word Segmentation. Cognition, 96,
233–262.
De Bree, E., Wijnen, F., and Zonneveld, W. (2006). Word Stress Production in Three-
Year-Old Children at Risk of Dyslexia. Journal of Research in Reading, 29(3), 304-317.
Findlay, J.M. (1978). Estimates on Probability Functions: A More Virulent PEST.
Perception and Psychophysics, 23, 181-185.
Forgeard, M., Gottfried, S., Norton, A., Rosam, C., Iyegar, U., and Winner, E. (2008).
The Relation Between Music and Phonological Processing in Normal-Reading Children
and Children with Dyslexia. Music Perception, 25 (4), 383-390.
Gallagher, A., Frith, U., and Snowling, M.J. (2000). Precursors of Literacy Delay Among
Children at Genetic Risk of Dyslexia. Journal of Child Psychology and Psychiatry, 41
(2), 202-213.
Ganschow, L., Lloyd-Jones, J., and Miles, T. R. (1994). Dyslexia and Musical Notation.
Annals of Dyslexia, 44, 185-202.
Gibson, L.Y., Hogben, J.H., and Fletcher, J. (2006). Visual and Auditory Processing and
Component Reading Skills in Developmental Dyslexia. Cognitive Neuropsychology,
23(4), 621-642.
Giraud, A., Lorenzi, C., Ashburner, J., Wable, J., Johnsrude, I., Frackowiak, R.S.J., and
Kleinschmidt, A. (2000). Representation of the temporal envelope of sounds in the
human brain. Journal of Neurophysiology, 84 (3), 1588–1598.
Goswami, U. (2003). How to Beat Dyslexia. The Psychologist, 16(9), 462-465.
Goswami, U., Thomson, J., Richardson, U., Stainthorp, R., Highes, D., Rosen, S., and
Scott, S.K. (2002). Amplitude Envelope Onsets and Developmental Dyslexia: A New
Hypothesis. PNAS, 99(16), 10911-10916.
Guttorm, T.K., Leppanen, P.H.T, Richardson, U, and Lyytinen, H. (2001). Event-Related
Potentials and Consonant Differentiation in Newborns with Familial Risk for Dyslexia.
Journal of Learning Disabilities, 34 (6), 534-544.
Hall, D.A., Haggard, M.P., Akeroyd, M.A., Summerfield, A.Q., Palmer, A.R., Elliot,
M.R., and Bowtell, R.W. (2000). Modulation and Task Effects in Auditory Processing
37
Measured using fMRI. Human Brain Mapping, 10, 107-119.
Halliday, L.F. and Bishop, D.V.M. (2006). Auditory Frequency Discrimination in
Children with Dyslexia. Journal of Research in Reading, 29(2), 213-228.
Hamalianen, J., Leppanen, P.H.T., Torppa, M., Muller, K., and Lyytinen, H. (2005).
Detection of Sound Rise Time by Adults with Dyslexia. Brain Language, 94(1), 32-42.
Heil, P. (2003). Coding of Temporal Onset Envelope in the Auditory System. Speech
Communication, 41, 123-134.
Heil, P. and Neubauer, H. (2001). Temporal Integration of Sound Pressure Determines
Thresholds of Auditory-Nerve Fibers. The Journal of Neuroscience, 21 (18), 7404-7415.
Heil, P. and Neubauer, H. (2003). A Unifying Basis of Auditory Thresholds Based on
Temporal Summation. PNAS, 100 (10), 6151-6156.
Heil, P. and Neubauer, H. (2004). Towards a Unifying Basis of Auditory Thresholds: The
Effects of Hearing Loss on Temporal Integration Reconsidered. Journal of the
Association for Research in Otolaryngology, 5 (4), 436-458.
Irvy, R.B. and Keele, S.W. (1989). Timing Functions in the Cerebellum. Journal of
Cognitive Neuroscience, 1 (2), 136-152.
Joris, P.X., Schreiner, C.E., and Rees, A. (2003). Neural Processing of Amplitude-
Modulated Sounds. Physiological Reviews, 84, 541-577.
Leppanen, P.H.T., Richardson, U., Pihko, E., Eklund, K.M., Guttorm, T.K., Aro, M., and
Lyytinen, H. (2002). Brain Responses to Changes in Speech Sound Durations Differ
Between Infants with and without Familial Risk for Dyslexia. Developmental
Neuropsychology, 22(1), 407-422.
Leppanen, P.H.T, Pihko, E., Eklund, K.M., and Lyytinen, H. (1999). Cortical Responses
of Infants with and without a Genetic Risk for Dyslexia: II. Group Effects. NeuroReport,
10 (5), 969-973.
Lorenzi, C., Dumont, A., and Fullgrabe, C. (2000). Use of Temporal Envelope Cues By
Children with Developmental Dyslexia. Journal of Speech, Language, and Hearing
Research, 43 (6), 1367-1379.
McArthur, G. M., and Bishop, D. V. M. (2001). Auditory Perceptual Processing in
People with Reading and Oral Language Impairments: Current Issues and
Recommendations. Dyslexia, 7, 150–170.
Menell, P., McAnally, K.I., and Stein, J.F. (1999). Amplitude Modulation in Adult
38
Dyslexic Listeners. Journal of Speech Language and Hearing Research, 42, 797-803.
Morris, R.D., Stuebing, K.K., Fletcher, J.M., Shaywitz, S.E., Lyon, G.R., Shankweiler,
D.P., Katz, L., Francis, D.J., and Shaywitz, B.A. (1998). Subtypes of Reading Disability:
Variability Around a Phonological Core. Journal of Educational Psychology, 90(3), 347-
373.
Muneaux, M., Ziegler, J.C., Truc, C., Thomson, J., and Goswami, U. (2004). Deficits in
Beat Perception and Dyslexia: Evidence from French. NeuroReport, 15(8), 1255-1259.
Overy, K. (2003). Dyslexia and Music. From Timing Deficits to Musical Intervention.
Annals of the New York Academy of Sciences, 999, 497-505.
Pasquini, E.S., Corriveau, K.H., and Goswami, U. (2007). Auditory Processing of
Amplitude Envelope Rise Time in Adults Diagnosed With Developmental Dyslexia.
Scientific Studies of Reading, 11 (3), 259-286.
Ramus, F., Rosen, S., Dakin, S.C., Day, B.L., Castellote, J.M., White, S., and Frith, U.
(2003). Theories of Developmental Dyslexia: Insights From a Multiple Case Study of
Dyslexic Adults. Brain, 126, 841-865.
Richardson, U., Thomson, J., Scott, S.K., and Goswami, U. (2004). Auditory Processing
Skills and Phonological Representation in Dyslexic Children. Dyslexia, 10, 215-233.
Rosen, S. (2003). Auditory processing in dyslexia and Specific Language Impairment: Is
There a Deficit? What is its Nature? Does it explain anything? Journal of Phonetics, 3
(4), 509–527.
Scott, S.K. (1998). The Point of P-Centres. Psychological Research, 61(1), 4-11.
Semel, E.M., Wiig, E.H., and Secord, W.A. (1995). Clinical Evaluation of Language
Fundamentals – Third Edition. San Antonio: Psychological Corporation.
Snowling, M.J. (2000). Dyslexia. Blackwell Publishers: Massachusetts.
Sperling, A., Lu, Z.L, Manis, F.R., and Seidenberg, M.S. (2005). Deficits in Perceptual
Noise Exclusion in Developmental Dyslexia. Nature Neuroscience, 8, 1-2.
Stanovich, K.E. and Siegel, L.S. (1994). Phenotypic Performance Profile of Children
with Reading Disabilities: A Regression-Based Test of the Phonological-Core Variable-
Difference Model. Journal of Educational Psychology, 86, 24–53.
Stark, R.E. and Tallal, P. (1981). Selection of Children with Specific Language Deficits.
Journal of Speech and Hearing Disorders, 46, 114-122.
39
Stein, J. (2001). The Magnocellular Theory of Developmental Dyslexia. Dyslexia, 7(1),
12-36.
Surányi, S., Csépe, V., Richardson, U., Thomson, M. J., Honbolygó, F., and Goswami, U.
(2006). Sensitivity to Rhythmic Parameters in Dyslexic Children: A Comparison of
Hungarian and English. Paper submitted for publication.
Tallal, P. (1980). Auditory Temporal Perception, Phonics and Reading Disabilities in
Children. Brain and Language, 9, 182–198.
Thomson, J., Fryer, B., Maltby, J., and Goswami, U. (2006). Auditory and Motor Rhythm
Awareness in Adults with Dyslexia. Journal of Research in Reading, 29(3), 334-348.
Thomson, J.M. and Goswami, U. (2008). Rhythmic Processing in Children with
Developmental Dyslexia: Auditory and Motor Rhythms Link to Reading and Spelling.
Journal of Physiology-Paris, 102 (1-3), 120-129.
Torgensen, J.K., Wagner, R.K., and Rashotte, C.A. (1999). Comprehensive Test of
Phonological Processing. Austin, TX: PRO-ED, Inc.
Torgensen, J.K., Wagner, R.K., and Rashotte, C.A. (1999). Test of Word Reading
Efficiency. Austin, TX: PRO-ED, Inc.
Vos, P.G., Mates, J., and van Kruysbergen, N.W., (1995). The Perceptual Centre of a
Stimulus as the Cue for Synchronization to a Metronome: Evidence from Asynchronies.
The Quarterly Journal of Experimental Psychology. A, Human Experimental Psychology,
48(4), 1024-40.
Wade-Woolley, L. and Wood, C. (2006). Editorial: Prosodic Sensitivity and Reading
Development. Journal of Research in Reading, 29 (3), 253-257.
Weschler, D. (1991). Weschler Intelligence Scale for Children – Third Edition. San
Diego, CA: The Psychological Corporation.
Whalley, K. and Hansen J. (2006). The Role of Prosodic Sensitivity in Children’s
Reading Development. Journal of Research in Reading, 29(3), 288-303.
Wiederholt, J.L. and Blalock, G. (2000). Gray Silent Reading Tests. Austin, TX: PRO-
ED, Inc.
Wiederholt, J.L and Bryant, B.R. (1992). Gray Oral Reading Tests. Austin, TX: PRO-
ED, Inc.
40
Wilcox, R.R. (2003). Applying Contemporary Statistical Techniques. Los Angeles, CA:
Academic Press.
Witton, C., Stein, J.F., Stoodley, C.J., Rosner, B.S., and Talcott, J.B. (2002). Separate
Influences of Acoustic AM and FM Sensitivity on the Phonological Decoding Skills of
Impaired and Normal Readers. Journal of Cognitive Neuroscience, 14(6), 866–874.
Wolf, M. and Bowers, P.G. (1999). The Double-Deficit Hypothesis for the
Developmental Dyslexias. Journal of Educational Psychology, 91(3), 415-438.
Wood, C. (2006). Metrical Stress Sensitivity in Young Children and its Relationship to
Phonological Awareness and Reading. Journal of Research in Reading, 29(3), 270-287.
Woodcock, R. W., McGrew, K.S., and Mather, N. (2001). Woodcock-Johnson III. Itasca,
IL: Riverside Publishing.

Abstract (if available)

Abstract A recent theory proposed by Goswami et al (2002), dubbed the P-Center hypothesis, posits that the characteristic cognitive and behavioral patterns observed in developmental dyslexia are a result of low-level auditory impairments. Previous studies have found that children with developmental dyslexia perform worse on P-center perception tasks when compared to chronological age matched controls, younger reading level controls have intermediate thresholds, and this deficit has also been observed in children with specific language impairment. The current study found similar trends in children aged between 7-16, in that children with dyslexia and dyslexia with additional language difficulties had less sensitive mean thresholds on the P-center perception tasks than chronological age matched controls. If these trends persist, then more ecologically valid stimuli and more sophisticated methods should be used to explore the P-center deficit in children with developmental dyslexia.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Examining the neuroanatomical correlates of reading in developmental dyslexia

Asset Metadata

Creator Beattie, Rachel Lynette (author)

Core Title P-center perception in children with developmental dyslexia: do low level auditory deficits underlie reading, spelling, and language impairments?

School College of Letters, Arts and Sciences

Degree Master of Arts

Degree Program Psychology

Publication Date 10/31/2008

Defense Date 10/05/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag auditory deficit,developmental dyslexia,OAI-PMH Harvest,P-center,rise time

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Manis, Franklin R. (committee chair), Farver, Jo Ann M. (committee member), Lu, Zhong-Lin (committee member)

Creator Email beattie@usc.edu,rachel.beattie@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1737

Unique identifier UC1412850

Identifier etd-Beattie-2463 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-116308 (legacy record id),usctheses-m1737 (legacy record id)

Legacy Identifier etd-Beattie-2463.pdf

Dmrecord 116308

Document Type Thesis

Rights Beattie, Rachel Lynette

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu