1. Auditory Processing Deficits in Dyslexia:
Task or Stimulus Related?
Karen Banai1,3
and Merav Ahissar2
1
Department of Neurobiology and 2
Department of Psychology
and Interdisciplinary Center for Neural Computation, Hebrew
University of Jerusalem, Jerusalem 91904, Israel 3
Current
address: Department of Communication Sciences and
Disorders, Northwestern University,
Evanston, IL 60208, USA
The nature of the fundamental deficit underlying reading disability is
the subject of a long-standing debate. We previously found that
dyslexics with additional learning difficulties (D-LDs) perform
poorly in simple auditory tasks. We now tried to determine whether
these deficits relate to stimulus or task complexity. We found that
the degree of impairment was dependent on task rather than
stimulus complexity. D-LDs could adequately detect and identify
mild frequency changes in simple pure tones and minimal phonemic
changes in complex speech sounds when task required only simple
same--different discriminations. However, when task required the
identification of the direction of frequency change or the ordinal
position of a repeated tonal or speech stimulus, D-LDs’ performance
substantially deteriorated. This behavioral pattern suggests that
D-LDs suffer from a similar type of deficits when processing
speech and nonspeech sounds. In both cases, the extent of diffi-
culties is determined by the structure of the task rather than by
stimulus composition or complexity.
Keywords: auditory processing, dyslexia, frequency discrimination,
perceptual memory, perceptual processing, working memory
Introduction
Dyslexics are individuals with substantial and continuous
reading difficulties in spite of adequate general intelligence
and education. They comprise ~10% of school children (Lyon
1995; Shaywitz 1998; Snow and others 1998). Though a sub-
population of dyslexics has particularly high general cognitive
abilities, enabling them to compensate for their poor reading,
most dyslexics have additional learning difficulties (Fletcher and
others 1994).
Most dyslexics, with and without additional learning difficul-
ties, suffer from poor phonological processing (Snowling 2000).
They have difficulties ‘‘hearing’’ words that are composed of
smaller speech segments and in ‘‘manipulating’’ speech sounds.
These impairments are directly linked to their reading difficulty
because decoding of the alphabetical script requires mapping
visual symbols to basic speech sounds (Snowling 2000).
Another typical characteristic of dyslexia is poor verbal
working memory (Siegel and Ryan 1989; Swanson 1993; Vargo
and others 1995). Working memory is the ability to retain a set of
stimuliwhileperformingadditional cognitive operationsonthem
(Baddeley 1986). Basic tasks, like decoding unfamiliar words and
simple arithmetic calculations, require holding parts (speech
segments or digits) while manipulating other parts of the input
stream. Consequently, poor working memory may impede the
performance in a broad range of academic tasks including, but
not specific to, reading (Gathercole and Pickering 2000).
A 3rd set of observations relates to dyslexics’ performance in
simple psychoacoustic tasks. Many studies (e.g., Tallal 1980; see
review by Wright and others 2000) reported dyslexics’ poor
performance in such tasks, though both the prevalence and the
functional role of these deficits have been heatedly debated
(Rosen 1999).
Recently, we found that poor psychoacoustic performance is
characteristic of a specific, large, subpopulation of dyslexics
with additional learning difficulties (D-LDs). These individuals
also suffer from particularly poor verbal working memory,
the extent of which is correlated with the degree of their
psychoacoustic and reading deficits (Amitay, Ben-Yehudah,
and others 2002; Banai and Ahissar 2004). Though D-LDs
perform standard problem-solving tasks within the normal
range of the general population, they have severe difficulties
performing tasks with heavy working memory load, and they
often need special academic support to graduate high school
(Banai and Ahissar 2004).
The relationships between D-LDs’ phonological, psycho-
acoustic, and working memory deficits are not clear. Because
human working memory has mostly been studied with phono-
logical material, it is hard to decipher whether the difficulty in
manipulating speech sounds stems from poor processing of
sounds—that is, a stimulus-specific deficit—or from a general
difficulty in interstimulus retention and manipulation—that is,
a task-specific deficit. Similarly, because adequate performance
on any psychoacoustic discrimination task requires both en-
coding of the specific stimuli to be discriminated and the dis-
crimination process itself, that is, the need to serially retain and
compare stimuli, when discrimination is impaired, it is hard to
dissociate whether poor performance results from a stimulus-
specific deficit (encoding auditory stimuli) or from a deficit
related to the discrimination task at hand (retention, compar-
ison, decision).
In the current study, we asked whether D-LDs’ deficits are
stimulus or task specific using both pure tones and complex
speech sounds. Specifically, we compared the predictions of
2 types of hypotheses regarding the mechanisms underlying
D-LDs’ poor performance on auditory discrimination tasks.
1. D-LDs have poor representations of auditory stimuli. Hence,
they will have difficulties discriminating between similar
auditory stimuli, regardless of task complexity. The extent of
their difficulties will therefore increase with the subtlety of
the required stimulus discrimination. This bottom-up hy-
pothesis is parsimonious and consistent with many previous
findings, most notably by Tallal (1980) and the following
behavioral (De Weirdt 1988; Reed 1989; Nicolson and others
1995; Hari and Kiesila 1996; Kraus and others 1996;
McAnally and Stein 1996; Adlard and Hazan 1998; Witton
and others 1998; Hari and others 1999; Heath and others
1999; Talcott and others 1999; Ahissar and others 2000;
Cerebral Cortex December 2006;16:1718--1728
doi:10.1093/cercor/bhj107
Advance Access publication January 11, 2006
Ó The Author 2006. Published by Oxford University Press. All rights reserved.
For permissions, please e-mail: journals.permissions@oxfordjournals.org
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

2. Waber and others 2000; Amitay, Ahissar, and Nelken 2002;
Ramus and others 2003) and electrophysiological (Baldeweg
and others 1999; Nagarajan and others 1999; Ahissar and
others 2001; Kujala and others 2003; Maurer and others
2003; Renvall and Hari 2003) studies. It predicts that D-LDs’
performance will be poor when confronted with a percep-
tual challenge, that is, when either high-resolution auditory
discriminations or complex auditory processing are required,
regardless of the behavioral task used.
2. D-LDs’ ability to perform auditory discriminations is limited
by task requirements (i.e., protocol of stimulus presentation
and required manipulations). Thus, the same type of stimuli
(speech or nonspeech) will induce more severe difficulties
when task complexity is increased (i.e., by increasing
memory load). Although quite a few previous studies as-
sessed psychoacoustic abilities in related populations, they
typically used a single behavioral paradigm (e.g., Ahissar and
others 2000; Amitay, Ahissar, and Nelken 2002; Ramus and
others 2003) and hence could not dissociate whether dif-
ficulties related to basic stimulus processing or to the re-
quired manipulations (though see France and others 2002).
To compare the predictions of these 2 hypotheses, we used 3
frequency discrimination tasks, using the same stimuli with
different degrees of task complexity, and 2 speech perception
tasks, again using the same speech tokens with 2 different
procedures.
Materials and Methods
Participants
The study was conducted in 2 test populations. The 1st was composed
of 57 8th grade students (all female, aged 14.2 ± 0.4) from 2 classes, 32
from a regular public school, and 25 from a private school for students
with mild learning disabilities (LDs). The control (regular) school was
chosen based on the similar demographic characteristics of the students
in the 2 schools. Two paradigms of frequency discrimination as well as
reading, memory, and cognitive measures were administered to this
population. Five students had to be removed from data analysis after all
assessments were completed: One from the LD class had generally poor
cognitive abilities, as measured by standard tests. For 1 control par-
ticipant, we could not measure a reliable threshold for any of the psy-
chophysical tasks, and we thus removed her data from further
analysis as well. Two other students in the control class were found to
be reading disabled, and their data were therefore also removed. Data of
another student from the control class were removed because of
a hearing impairment. We thus report the results of 24 students from
the special school and 28 normal learning control students.
The 2nd population comprised 35 7th grade students (all female, aged
13.1 ± 0.4) from the same school for mildly learning-impaired individ-
uals, who participated in this study as a part of a larger research program
in which they took part. Data of 1 student suspected of a hearing loss
were removed from the analysis, and thus, data of 34 7th grade students
are reported.
The parents of all participating students gave their consent after
receiving letters with information regarding the study. All participants
completed all tasks, unless otherwise noted in the text.
Group Classification
None of the students from the regular school had a history of learning
problems, and they were thus defined as a normal-learning control
group (control I). Students from the LD school all had a history of LD,
reading related or otherwise. For the purpose of the current study, they
were divided to 2 groups: one with dyslexia (D-LD) and the other with
normal reading and phonological abilities, which served as a same-
classroom control group (control II). Classification was based on 2
measures that should be impaired among dyslexics based on current
definitions stressing the significance of phonological deficits to the
diagnosis of dyslexia (Lyon 1995; Snowling 2000): reading of pseu-
dowords and phonological awareness (see task descriptions sub-
sequently). Based on these 2 measures, we calculated a combined
phonological score with respect to normal-learning controls (control I
average without the 2 individuals with reading disability). LDs whose
phonological score was lower than the controls’ average by 1.5 standard
deviation or more were defined as dyslexics (n = 15 in the 8th grade and
22 in the 7th grade). The rest, whose phonological abilities were within
the range of the controls, were defined as same-classroom controls
(control II; n = 9 in the 8th grade and 12 in the 7th grade).
Cognitive Performance
Unlike discrepancy-based definitions of dyslexia, current diagnostic
criteria, which were used in this study (Siegel 1992; Waber and others
2000), do not require a discrepancy between reading abilities and
intelligence quotient (IQ) scores but rather within the normal IQ ( >80).
Although we could not conduct a full IQ test during the study, all
students in the special school were administered a detailed psycho-
educational assessment before being admitted to the school to ensure
that they have the potential to graduate with a full high school diploma
by the end of the 12th grade. This testing rules out the possibility of
IQ < 80 among the students (except for 1 student, whose data were
excluded, as explained earlier). In addition, the Raven standard pro-
gressive matrices scores (Raven and others 2000) of all these students
were >35, which excludes the lower 25% of the population.
Materials
Psychophysical Tasks
All stimuli were presented to both ears through Sennheiser HD-265
linear headphones and a TDT system III signal generator (Tucker Davis
Technologies, Alachua, FL) controlled by a PC in a quiet room in each
school.
Auditory frequency discrimination. Participants made frequency
discriminations in 3 conditions. The frequency of the test (higher)
tone was always changed in a 2 down/1 up staircase procedure (step
size was 40 Hz for the first 5 reversals and 5 Hz thereafter) converging at
a performance level of 70.7% (Levitt 1971). The frequency of the
reference tone was fixed, 1000 Hz. A pleasant visual feedback was
provided for correct responses.
1. High--low discrimination (Hi/Lo): Two 50-ms tones with 1000-ms
interstimulus interval (ISI) were presented in each trial. The
students had to indicate which tone was higher (1st or 2nd).
Assessment terminated after 50 trials or 10 reversals. Thresholds
were measured twice in the 8th grade, at the beginning and at the
end of the 1st session, and the average just noticeable difference
(JND) was used in the statistical analysis. Due to the high correlation
between the two, only 1 assessment was conducted in the 7th grade.
2. Same--different discrimination (2-tone S/D): Two 50-ms tones (with
1000-ms ISI) were presented in each trial. The listener had to
indicate whether they were same or different. Same and different
trials had an equal probability of appearance. Assessment terminated
after 100 trials or 10 reversals (based only on trials with different
stimuli). Threshold was measured once in the 2nd testing session.
3. Same--different discrimination with 3 tones (3-tone S/D): a fixed
reference tone was followed by a 1500-ms silent interval and then by
2 other tones (with 950-ms interval between them), one of which
was a repetition of the reference and the other was different.
Participants had to indicate which tone (2nd or 3rd) was the same as
the reference.
Participants were given a practice block of 15 trials (with a frequency
difference of 1000 Hz between the tones) to familiarize them with the
task prior to assessment. Initial frequency difference in the test blocks
was 500 Hz. If they had 2 errors or more, they were given another
practice block. If they still had 2 errors or more, testing begun but with
Cerebral Cortex December 2006, V 16 N 12 1719
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

3. a larger initial frequency difference, 800 Hz. Participants responded
to the stimuli in each trial orally, by telling their answer to the
experimenter who immediately entered it to the computer to continue
the procedure. This mode of response was chosen to minimize errors
due to sensory-motor confusions. This interactive mode further ensured
that all participants were maximally alert throughout the assessment
procedure. However, because it introduced an intervening oral re-
sponses (within the sequence of pure tones) and somewhat prolonged
the assessment process, it may have elevated the estimated discrimina-
tion thresholds.
Calculation of frequency discrimination thresholds. Frequency JNDs,
that is, the minimal frequency difference reliably discriminated, were
used as indices of discrimination. For each participant, JNDs were cal-
culated based on the average of her last 5 reversals. Hi/Lo JNDs were
obtained for 54 8th grade and all 7th grade students (an equipment
failure occurred during the assessment of one of the LD students). Two
control students failed to achieve a reliable JND in the 1st assessment
(they had no reversals in the last 20 trials), and we thus used only their
2nd thresholds. Given the high correlation between the 2 Hi/Lo
measurements (r = 0.79, P < 0.001), we used the average JND for the
rest of the students in the 8th grade. S/D JNDs were obtained for 51 8th
grade (28 Controls I, 8 Controls II, and 15 D-LDs) and all 7th grade
participants.
Auditory duration discrimination. Seventh grade participants heard 2
tones and had to indicate which one was longer, the 1st or the 2nd.
Fixed reference duration was 100 ms in 1 condition and 400 ms in the
other. Step sizes were 40, 25, and 5 ms in the 100-ms condition and 40,
25, and 5 ms in the 400-ms condition. Step size changed after 3 reversals.
All other aspects of the task were identical to the Hi/Lo frequency
discrimination task.
Speech discrimination and identification. Eighth grade participants
were required to discriminate minimal phonemic pairs. The stimuli
were pairs of 2-syllable pseudowords differing in 1 consonant and
identical otherwise. The following consonant pairs were used: /d/-/t/,
/b/-/d/, /b/-/p/, /v/-/f/, /m/-/n/, and /s/-/z/, appearing either at the
beginning or at the end of the word. Between trials, 4 different syllables
were used (/a/, /e/, /i/, and /u/). All stimuli were produced by a native
female Hebrew speaker. Twenty-four different pairs were used.
The task had 2 conditions with 48 trials in each. 1) Two-word S/D
condition in which the 24 different pairs were supplemented with 24
same pairs. 2) Three-word S/D condition in which the participant first
heard a pair of pseudowords. After a 2-s interval, one of the pseudo-
words was repeated, and the task was to indicate which word it
was—the 1st or the 2nd. Each pair was repeated twice, each time with
a different repeated word.
Speech perception in noise. Minimal intensity level for accurate speech
perception in noisy background (narrow-band amplitude--modulated
noise) was tested using the same set of 2-syllable pseudowords de-
scribed earlier, normalized for intensity and duration. Total noise RMS
was 83-dB SPL (for fuller description, see Putter-Katz and others 2005).
On each trial, a single pseudoword was played and participants were
asked to accurately repeat it. The intensity of the pseudowords was
adapted in a 3 down/1 up staircase procedure (converging on ~80%
correct) while noise level remained fixed. The assessment terminated
after 85 trials or 15 reversals. Identification threshold was the arithmetic
mean of the last 10 reversals. This assessment was conducted in the
7th grade.
Reading
Oral reading of pointed Hebrew words and pointed nonwords was
assessed using a list of 24 pseudowords followed by a list of 24 words
(see Ben-Yehudah and others 2001). The pseudowords were designed
to mimic the real Hebrew words in the word list. These reading tests,
designed by A. Deutsch (for details, see Deutsch and Bentin 1996), were
previously found to be reliable indicators of reading difficulties. Subjects
had to read aloud, as accurate and as fast as possible from the printed
lists presented to them. Both accuracy and speed were recorded.
Phonological Awareness
Phonological awareness was assessed using the Hebrew version of the
spoonerism task (developed by Peleg and Ben Dror, unpublished test
materials), consisting of phoneme deletion and replacement. Listeners
were orally presented with a word pair and were asked to swap the 1st
phonemes of the 2 words (e.g., white pig / pite wig). In the process of
introducing the task, all participants successfully performed its simple
substeps (phoneme deletion/replacement). Twenty word pairs were
presented.
Memory Tests
Verbal memory was assessed using Digit Span, a subtest of Wechsler
intelligence scale for children (WISC–R95 Wechsler 1998). This test
contains 2 parts. In the 1st part, participants are asked to repeat a list of
orally presented digits in the order they were presented. The maximal
length of successfully reproduced sequences characterizes the capacity
or span of verbal memory. In the 2nd part, subjects are asked to repeat
the sequence of digits in the reverse order. The 2nd part is more
complex and measures the combined effect of capacity and manipula-
tion in verbal working memory. This test was scored using the Wechsler
scaled score.
A Hebrew version of the ‘‘sentence span’’ task adapted by Ben Dror
and Shani (1998) from the reading span task originally developed by
Daneman and Carpenter (1980) was used to assess retention of words
through interfering stimuli and task switching. Seventh grade partic-
ipants were asked to listen to a series of orally presented short sentences
with a missing last word, complete the last word of each sentence, and
recall all the last words in the order of completion at the end of the
series. Series length increases from 2 to 6 sentences. Scoring emphasizes
both the number and the order of correctly recalled words.
Nonverbal Cognitive Abilities
Block design (WISC-R95, Israeli version) was used as a measure of
general cognitive abilities. Administration and scoring followed the
standard procedure described in the Administration and Scoring manual
of the test.
Results
Population Profile
Table 1 shows the average scores of the 3 groups of 8th grade
participants: control I, control II, and D-LDs in reading and
reading-related tests and in standard cognitive tasks. Table 2
shows reading, reading-related, and standard cognitive scores of
the 7th grade participants. As expected, D-LDs’ performance
was particularly impaired in the reading and phonological tasks
compared with both control groups.
General Cognitive Abilities
The 2 control groups did not differ significantly on any of the
reading or cognitive measures administered. Thus, nondyslexic
LDs (control II) are similar to normal-learning students in
general cognitive abilities and yet form a more adequate control
group to D-LDs in terms of having the same educational
background and school experience. D-LDs scored lower than
control I on block design, a standard measure of general
cognitive ability. However, their scores were not significantly
lower than those of control II (for 8th graders: t = 1.2, P > 0.22,
not corrected, to allow minimum chance of retaining the null
hypothesis that the groups do not differ cognitive tasks; for the
7th grade, see Table 2), whose scores were intermediate.
Moreover, D-LDs’ scores on this task were well within the
normal range. Table 2 further shows that scores in the 2 tasks
that often serve as a short version for measuring general IQ,
1720 Auditory Deficits in Dyslexia: Task or Stimulus d
Banai and Ahissar
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

4. vocabulary, and block design (Woerner and Overstreet 1999)
did not significantly differ between D-LDs and control II.
Processing Nonspeech Stimuli
Frequency Discrimination
Figure 1 shows average discrimination thresholds (JNDs) of the
3 subject groups on the S/D and the Hi/Lo (1st and 2nd
assessments) frequency discrimination tasks. A clear difference
in thresholds is evident. The Hi/Lo task yielded higher JNDs
than the S/D task (Ftask = 19.2, P < 0.001). However, this
difference was the result of D-LDs having huge difficulties with
the Hi/Lo task (in both assessments) compared with the 2 other
groups. The 2 control groups, on the other hand, did not differ
from each other. The average of the 2 measurements was 22 ±
18%, 16 ± 10%, and 52 ± 28% in the control I, control II, and
D-LD groups, respectively, with significant group (F = 9.7,
P < 0.001) and interaction (F = 6.9, P = 0.002) effects; the 2
post hoc comparisons between D-LDs and the other groups
were significant at P < 0.005, whereas those between the 2
control groups were not significant. Percent success around
these thresholds did not differ significantly between the 3
groups (81 ± 10%, 83 ± 6%, and 74 ± 17% in the control, LD, and
D-LD groups, respectively; F = 1.7, P > 0.18), indicating that
group differences do not stem from measuring different levels
of performance accuracy.
On the other hand, on the S/D task, neither performance level
(F = 0.38, P = 0.687) nor performance accuracy (average percent
correct was 84 ±6, 84 ±4 and 82 ±6 for Controls I, Controls II and
D-LDs, respectively) significantly differed between the 3 groups.
Comparing performance in these 2 frequency discrimination
tasks on a subject by subject basis, we found no correlation
between thresholds (r = 0.09, P > 0.5 in the entire 8th grade
group), indicating that different bottlenecks may limit perfor-
mance in these 2 tasks. Figure 2 illustrates that performance of
Table 2
Group means (SD) in reading-related and standard tasks: 7th grade
Control II (n 5 12) D-LD (n 5 22) ta
P
Reading Accuracyb
Nonwords 91.3 (7) 57.6 (14) 7.6 0.001
Words 97.2 (3) 86.2 (7) 5.1 0.001
Reading speedc
Nonwords 48 (11) 35 (12) 3.2 0.003
Words 100 (29) 74 (26) 2.6 0.013
Phonology
Spoonerismb
83 (16) 52 (23) 3.7 0.001
Standard scoresd
Block design 11.1 (2) 9.5 (3) 1.7 0.092
Digit span 10.2 (2) 7.7 (2) 3.8 0.001
Vocabulary 7.6 (2) 7.3 (2) 0.5 0.633
Sentence spane
32 (5) 31 (4) 0.59 0.562
a
t and P based on t-test between D-LDs and LDs.
b
Percent correct.
c
Words/min.
d
Wechsler scaled scores.
e
Score.
Hi/Lo1 Hi/Lo2 2-tone S/D
0
10
20
30
40
50
60
70
JND(%)
Figure 1. Frequency discrimination—Hi/Lo versus S/D. Frequency JNDs of D-LD
(filled bars), control II (empty bars with thick lines), and control I (empty bars with thin
lines) participants on the Hi/Lo (2 left sets of bars) and S/D (right bar set) tasks. Hi/Lo1
and Hi/Lo2 denote the 1st and 2nd assessments of this condition, respectively. As
explained in the text, average of the 2 measurements was used for statistical analysis.
Data are from 8th grade participants. Error bars denote ± 1 standard error of the mean.
0 25 50 75 100
0 25 50 75 100
0
4
8
12
0
4
8
12
VerbalSpan
0 25 50 75 100
0
25
50
75
100
0
25
50
75
100
Hi/Lo JND (%)
PhonAwareness(%corr)
0 25 50 75 100
S/D JND (%)
Figure 2. Frequency discrimination versus verbal memory and phonological
awareness among D-LDs. Hi/Lo JNDs (average of the 2 assessments, left panels)
are significantly correlated with both verbal memory (measured with digit span) and
phonological awareness (measured with spoonerism) scores; 2-tone S/D JNDs (right
panels) are not. Data are shown from the 8th grade.
Table 1
Group means (SD) in reading-related and standard tasks: 8th grade
Control I
(n 5 28)
Control II
(n 5 9)
D-LD
(n 5 15)
Fa
P
Reading accuracyb
Nonwords 89.5 (10)*** 84.7 (11)*** 52.8 (15) 48.6 0.001
Words 98.1 (3)*** 97.7 (4)*** 83.6 (11) 27.7 0.001
Reading Speedc
Nonwords 51 (14) 47 (12)* 38 (16) 4.1 0.02
Words 90 (27) 96 (28) 78 (25) 1.6 0.22
Phonology
Spoonerismb
88 (11)*** 85 (19)*** 50 (30) 18.7 0.001
Standard scoresd
Block design 11.7 (2)** 10.7 (3) 8.8 (3) 6.3 0.004
Digit span 10.2 (3)** 10.3 (2)* 7.7 (2) 8.1 0.001
a
F and P are the results of a 1-way analysis of variance between the 3 groups.
b
Percent correct.
c
Words/min.
d
Wechsler scaled scores.
*P 0.05, **P 0.01, ***P 0.001 on a post hoc test between D-LDs and each of
the other groups.
Cerebral Cortex December 2006, V 16 N 12 1721
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

5. the D-LD participants in the Hi/Lo (left), but not in the S/D task
(right), was significantly correlated with both their verbal
memory (i.e., digit span: r = 0.68, P = 0.005) and phonological
awareness scores (i.e., spoonerism: r = 0.82, P < 0.001), 2
measures known to be causally linked with reading skills. A
similar correlation pattern was observed among LDs (not
shown), though it was only marginally significant due to the
smaller size of this group (r = 0.72 and 0.55, P = 0.04 and 0.15,
respectively, for digit span and spoonerism). Note that S/D JNDs
are not correlated with these measures even though they are
quite scattered (the range of S/D JNDs among controls is
0.5--40%, i.e., there is no floor effect).
Duration Discrimination
A plausible explanation for D-LDs’ difficulties in the Hi/Lo task
relates to the explicit analogy between pitch and height, which
is required by this task (participants were asked ‘‘which tone
was higher?’’) and could be particularly difficult for D-LDs. In
that case, their impaired performance may stem from the
specific requirement embedded in frequency discriminations
rather than from the general requirement of stimulus evalua-
tion. To assess whether this is indeed the case, we designed
another task, using the same behavioral paradigm as in the
Hi/Lo discrimination task; however, comparisons referred to an
orthogonal dimension—stimulus duration. Stimulus duration
discrimination may be a more attention-demanding task in
terms of requiring continuous monitoring throughout stimulus
presentation and, thus, perhaps requires more complex neural
circuitry (Leon and Shadlen 2003). Yet, duration discrimination
(which tone is longer?) is an intuitive and straightforward
evaluation, which was immediately understood by all partic-
ipants, and was not susceptible to an interpretation that D-LDs
might not have fully understood the task.
We administered the 2 duration discrimination conditions to
the 7th grade test group (who showed the same behavioral pat-
tern on the Hi/Lo task as shown below). As shown in Figure 3,
D-LDs (n = 19), who had difficulties in performing Hi/Lo
frequency discrimination, had similar difficulties in performing
long/short duration discriminations around both durations,
compared with the control II group (n = 9), with no difficulties
in Hi/Lo discriminations (t = 2.9, P = 0.007 for 400-ms reference;
t = 2.5, P = 0.016 for the 100-ms reference). Performance
accuracy did not differ between the groups in either condition
(percent correct in the control II and D-LD groups were 76 ± 4%
and 75 ± 4% with 100-ms reference and 73 ± 4% and 73 ± 6%
with the 400-ms reference, respectively), indicating that differ-
ences in thresholds did not stem from measuring different levels
of performance accuracy.
Taken together, we have shown that the difficulties of our
dyslexic participants are not specific to the compared di-
mension. When subjects had to estimate the direction of
change, whether higher or longer, D-LDs had difficulties.
Order Judgments
Participants could adequately perform the S/D task by identi-
fying repetition of the stimulus without having to associate
stimulus position (1st or 2nd) with the stimulus type (higher or
lower). Though this task clearly requires some form of memory
trace, such a trace may be implicit and result from stimulus-
specific adaptation mechanisms. These mechanisms yield
smaller responses to previously presented stimuli and thereby
a greater response to novel stimuli, regardless of ‘‘how’’ they are
novel. These perceptual mechanisms thus signal novelty but are
not sufficient for determining its nature as required in the Hi/Lo
task. An example for such a mechanism at the level of the
auditory cortex is the mismatch negativity (MMN) response,
which signals change in a repetitive sequence of auditory
stimuli (see, e.g., Naatanen and others 2001).
Under S/D conditions, D-LDs had no difficulties, whereas in
the Hi/Lo task, they had substantial difficulties. The difference
between these tasks may thus stem from the Hi/Lo task being
generally more taxing for explicit memory mechanisms or
perhaps from its more specific requirement to assess the nature
of the difference between the stimuli. To resolve this question,
we administered an extended S/D task, using 3 stimuli. In this
task, a repetition always occurs, and the listener needs to
identify the ordinal position of the repeated stimulus. Moreover,
it requires retention through interfering stimuli and often for
longer durations (though duration per se is probably not
overtaxing for D-LDs, Ben-Yehudah and others 2004). Thus,
from a general working memory perspective, it is a taxing task,
yet no parametric evaluations are required. Therefore, if D-LDs’
difficulties stem from the requirement for parametric stimulus
evaluations of the Hi/Lo task, then they are not expected to
have difficulties with this task, even though it is demanding
from a working memory perspective.
We administered all 3 frequency discrimination tasks to the
7th grade group described earlier. In therepeated tasks (Figure 4,
2 left bars), we replicated our findings (illustrated in Fig. 1) that
D-LDs were impaired in Hi/Lo (their average was 37 ± 26%
compared with 10 ± 11% in the control II group; t = 4.8,
P < 0.001) but not in S/D frequency discrimination (10 ± 10%
vs. 8 ± 7%, respectively; t = 1, P > 0.3). In the novel 3-tone S/D
task, D-LDs were significantly poorer than LDs (19 ± 15% vs.
8 ± 10%, respectively; t = 3.1, P = 0.004; Fig. 4, rightmost bars),
indicating that even S/D comparisons are difficult for them
when accurate retention and comparison of previous stimuli is
required through subsequent intervening stimuli.
100 ms 400 ms
0
20
40
60
80
100
Reference Duration
JND(%)
Figure 3. Duration discrimination. Average duration JNDs of D-LDs (filled bars) and
controls II (open bars) with 100-ms (left) and 400-ms (right) reference tones (data are
from the 7th grade).
1722 Auditory Deficits in Dyslexia: Task or Stimulus d
Banai and Ahissar
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

6. Taken together, these data show that D-LDs have difficulties
in tasks that require explicit retention and comparison of series
of simple tones.
Processing Speech Elements
Comparing Speech Sounds
We now asked whether a similar pattern of difficulties would
characterize D-LDs’ performance in analogous tasks that use
complex speech sounds rather than simple tones as test stimuli.
Two conditions of a speech discrimination task (described
earlier) were administered to the 8th grade D-LD and control II
participants: simple 2-stimuli S/D task (15 D-LDs and 9 controls
II completed this condition) and 3-stimuli order judgment
S/D condition (14 D-LDs and 7 controls II completed this
task), similar to the 3-tone S/D frequency discrimination task.
Figure 5A shows average scores in these 2 tasks, illustrating that
D-LDs were significantly poorer than their peers in discrimi-
nating phonological components when the task required
judging the ordinal position of the pseudowords but not in
the simple 2-stimuli S/D condition (Ftask = 9.6, P = 0.006; Fgroup =
5.2, P = 0.05; the interaction term approached significance
Finteraction = 3.5, P = 0.078).
Similar to the findings with pure tones, the degree of deficit in
processing complex speech sounds was thus also dependent on
the task: Impairment was revealed only when a 3rd stimulus was
introduced. Furthermore, as shown in Figure 6, D-LDs’ perfor-
mance in the 3-pseudoword speech task, but not in the 2-
pseudoword task (not shown), was highly correlated with their
Hi/Lo frequency JNDs (r = 0.72, P = 0.003). These findings
suggest that the nature of D-LDs’ deficits may be similar for
complex speech components and simple pure tones.
Speech Perception In Noise
Our findings that S/D comparisons of speech components were
not difficult for D-LD individuals suggest that they do not have
severe difficulties in speech perception. Yet, given that these
stimuli were not very difficult (natural speech presented at
a comfortable rate in quiet) and most subjects got most pairs
right, more subtle speech perception deficits may have been
overlooked due to a ceiling effect. In order to ensure that it was
not the relative simplicity of the stimuli we used, we adminis-
tered an adaptive test of speech perception in noise (with the
same pseudowords used for the discrimination tasks but with
superimposed noise). In this task (originally described by
Putter-Katz and others 2005), the stimuli were very complex,
but the memory and cognitive loads were very low because
Hi/Lo 2-tone S/D 3-tone S/D
0
10
20
30
40
50
JND(%)
Figure 4. Frequency discrimination (7th grade). From left—Hi/Lo, 2-tone S/D, and
3-tone S/D frequency JNDs of D-LDs (filled bars) and controls II (open bars).
3-word SD 2-word SD
70
75
80
85
90
95
100
%correct
speech in noise
10
20
30
40
50
60
70
Identificationthreshold(dB-SPL)
A B
Figure 5. (A) Comparison of speech elements. Average performance of D-LDs (filled
bars) and controls II (open bars) on a 3-pseudoword S/D task (left) and on a
2-pseudoword S/D task (right). A significant intergroup difference is found only for the
3-element tasks. Data are from 8th grade participants. (B) Speech perception in
noise. Minimal intensity levels of D-LDs and LDs for speech perception in noise. Data
are from 7th grade participants.
0 25 50 75 100
70
80
90
100
Hi/Lo JND (%)
3-wordS/D(%corr)
Figure 6. Verbal versus nonverbal performance of D-LDs. JNDs in the Hi/Lo frequency
task and scores (percent correct) in the 3-pseudoword S/D task are highly correlated
(data are from the 8th grade students).
Cerebral Cortex December 2006, V 16 N 12 1723
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

7. subjects were only asked to repeat a single bisyllabic pseudo-
word. The minimal signal level needed to achieve 80% correct
repetition does not significantly differ between control II and D-
LDs (t = 1.7, P > 0.10), as shown in Figure 5B. The finding that
increasing signal complexity does not increase D-LD’s difficul-
ties indicates that basic perceptual mechanisms underlying
immediate auditory perception are adequate in these individu-
als. Moreover, speech identification thresholds were not corre-
lated with Hi/Lo JNDs among either D-LDs or control II (r = 0.03
and 0.14, respectively, P > 0.59), suggesting that different
mechanisms affect performance in these 2 tasks.
Do D-LDs Suffer from a Generalized Cognitive Deficit?
Because it was not possible to fully match D-LDs and the other
groups on measures of cognitive ability, it may be claimed that
D-LDs’ deficits in performing complex discriminations (Hi/Lo,
short/long, and 3-stimuli S/D comparisons) arise simply because
these tasks have a generally greater cognitive load than those
they performed adequately (2-stimuli S/D, a single pseudoword
repetition). Thus, their deficit may reflect a general cognitive
deficit rather than a task-specific one and is therefore not
surprising. For several reasons addressed subsequently, we
would like to suggest that this is not the case.
First, where the group differences on Hi/Lo frequency
discrimination related to group differences in IQ, we would
expect a correlation between block design and Hi/Lo frequency
JNDs. This was not the case. No significant correlations be-
tween block design and Hi/Lo frequency discrimination were
observed in either group (r = 0.2, 0.002, and 0.1 for control I
[n = 29], control II [n = 19], and D-LDs [n = 37], respectively,
P > 0.29; data pooled from 7th and 8th grades). Furthermore,
repeating the statistical analysis of the Hi/Lo and S/D frequency
discrimination tasks with block design scores as a covariate did
not change any of our findings, as expected from the lack of
correlation between block design and frequency discrimina-
tion: Main effects for group (F = 9.5, P < 0.001), task (Hi/Lo vs. S/
D: F = 5.33, P = 0.02), and a significant task 3 group interaction
(F = 3.88, P = 0.028) were observed, whereas the block design 3
task interaction (F = 1.43, P > 0.25) was not significant,
suggesting that the findings cannot be fully attributed to IQ
scores.
Second, as shown in Table 2, D-LDs could perform complex
cognitive tasks like sentence span adequately. We chose this
task because it is cognitively demanding in similar ways to our
previous tasks, that is, it requires information retention through
interfering stimuli, and taxes attention and memory in terms of
double load of sentence completion as well as retention. Yet, it
does not require stimulus manipulation, and retention can be
enhanced using the semantic content of the words and
sentences.
Taken together, we conclude that D-LDs’ unique pattern of
deficits in handling auditory stimuli is not a consequence of
a general factor limiting their performance in all cognitively
demanding tasks (for further discussion, see Banai and Ahissar
2004, 2005).
Discussion
Summary of Results
In this study, we found that D-LDs’ performance was adequate
in tasks that require identification or simple S/D discrimination
of either pure tones or complex speech sounds. On the other
hand, given exactly the same physical stimuli, D-LDs had
substantial difficulties in tasks that required either parametric
comparisons (Hi/Lo and long/short) or judging the ordinal
position of the repeated stimulus (whether a tone or a pseu-
doword). Moreover, high correlations were found between the
degree of difficulty in performance of complex tasks using
simple tones and speech components. These findings suggest
that the mechanisms underlying D-LDs’ impaired ability to
perform these demanding tasks are common to the 2 types of
stimuli. This common impairment is not in reasoning or general
cognitive abilities, as indicated by adequate performance on
a cognitively demanding, yet largely semantic based, task. We
thus propose that D-LDs’ deficits involve mechanisms of
operating on perceptual aspects of recently processed stimuli,
broadly termed working memory mechanisms.
Sex Differences in Dyslexia
All our participants were females because our study was
conducted in an all-female school. Thus, although our results
are ‘‘clean’’ in the sense that we report within-gender perfor-
mance and variability (see Liederman and others 2005), we
cannot directly generalize our findings to the male population.
The issue of prevalence of perceptual impairments among male
versus female dyslexics had not been systematically studied.
Moreover, its assessment would be complex because it should
also address related issues of reasoning abilities, language
impairments, and attentional deficits. At present, even the
relative prevalence of dyslexia among males and females is still
disputed (Shaywitz and others 1990; Share and Silva 2003).
An informal examination of our own accumulated data (~200
dyslexic and ~200 control participants) suggests no consistent
gender-related perceptual differences.
The Underlying Auditory Deficit of D-LD Individuals:
Accuracy of Representation or Stimulus Manipulation?
What is it that differentiates the tasks that posed particular
difficulties to D-LDs from those that did not? Tasks that pose
a great challenge for D-LDs could, in principle, be tasks that
require more refined auditory representations. Alternatively,
they might be generally more demanding tasks, requiring better
reasoning or decision-making abilities. Third, these tasks may
pound on specific stimulus operations that are selectively
impaired among D-LDs. We argue for the 3rd alternative, even
though the range of tasks that we applied is not sufficient for
accurate characterization of these operations.
The 1st interpretation can be refuted because increasing
stimulus complexity and hence processing requirements to
speech stimuli (with and without noise) did not specifically
hamper D-LDs’ performance. The ‘‘difficulty’’ interpretation can
be refuted for the following 3 reasons. First, levels of difficulty,
as measured by performance accuracy, were equated across
tasks and conditions by applying an adaptive procedure that
converged to the same relative level of performance. Second, as
shown in Figures 1 and 4, in both control groups, thresholds did
not differ between the Hi/Lo and the same--different conditions.
These tasks were just not more difficult for them. Third,
difficulty per se was not a sufficient impediment to D-LDs, as
shown by their adequate performance of the cognitively
demanding task of sentence span.
1724 Auditory Deficits in Dyslexia: Task or Stimulus d
Banai and Ahissar
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

8. What could be the specific operations that are impaired
among D-LDs? Although our set of paradigms can refute the
general cognitive or the perceptual difficulty hypotheses, they
leave many alternatives unresolved. One is that some aspects of
working memory mechanisms are impaired among D-LDs
because the tasks that were specifically impaired among D-
LDs put a heavier load on working memory. Our previous
studies refuted a retention deficit, that is, faster decay of
memory trace interpretation, because manipulating ISIs did
not affect D-LDs’ deficits (Ben-Yehudah and others 2004). Yet,
whether it is the buildup of stimulus-specific memory traces or
the operation of stimulus comparison that is impaired is beyond
the scope of this study. Both may be more taxing in the tasks
that posed greater difficulties to D-LDs.
Our finding that the behavioral paradigm used for assessment
is crucial to the degree of deficits revealed by dyslexics is
consistent with reports of previous studies both in the auditory
(e.g., Heath and others 1999; France and others 2002) and in the
visual domain (Ben-Yehudah and others 2001). However, most
studies applied only a single behavioral paradigm for evaluating
thresholds, and thus, the substantial interstudy variability was
mainly interpreted in terms of different population samples
rather than different behavioral tasks (see Discussion in Banai
and Ahissar 2004). An important lesson from our study is that
in order to decipher whether it is the task or stimuli that
taxes individuals, one needs to assess performance using at
least 2 behavioral tasks that greatly differ in their working
memory load.
Task-Specific Difficulties and Theories of Dyslexia
According to the dominant ‘‘core phonological deficit hypoth-
esis’’ of dyslexia (Snowling 2000; Vellutino and others 2004),
reading disability is a result of a specific deficit in phonological
processing. By this account, phonological deficits contribute
to the reading deficits, whereas deficits in the processing of
nonverbal stimuli stem from a separate source and are irrelevant
to reading. Data from the current study does not provide
support for a speech-specific deficit. Rather, our findings,
coupled with recent findings from imaging studies (LoCasto
and others 2004; Burton and others 2005; Luo and others 2005),
suggest that processing of both speech and nonspeech stimuli
shares at least some common processing that is impaired among
D-LDs. Thus, although deficits in nonverbal auditory processing
may not be directly relevant to reading, it is part of the same
underlying deficit contributing to the phonological processing
deficit and consequently poor reading.
The current finding that the degree of discrimination deficit
among D-LDs is task dependent and that D-LDs do not have
difficulties detecting speech in noise suggests that their auditory
processing is not impaired under all circumstances. However,
whether we define the deficits we found as ‘‘perceptual’’ (Tallal
and others 1993) or ‘‘postperceptual’’ depends on our terminol-
ogy of perception. We believe that the ‘‘auditory experience’’
(i.e., what one perceives) is different for D-LDs under some
conditions, depending on stimulus protocol and behavioral task.
Thus, an integrative view of perception as a dynamic, large-scale
process, which underlies this experience (Hochstein and
Ahissar 2002), attributes these deficits to perceptual processes.
This means that somewhere in the circuitry underlying our
immediate explicit perception, D-LDs’ processing is impaired.
However, the cortical circuitry underlying these processes is
broad and includes, among others, frontal areas, which are major
contributors to the retention and comparison aspects required
for adequate performance in most tasks.
Relation to Previous Imaging and Electrophysiological
Studies
Based on the strong dependence of D-LDs’ manifested deficits
on stimulus protocol and on task demands, we suggest that their
fundamental deficit does not reside in auditory cortex. This
suggestion is based on both lesion and imaging results. Neuro-
psychological studies suggest that a pervasive lesion of the
primary auditory cortex impairs S/D discriminations for both
simple tones and speech sounds (Tramo and others 2002), in
contrast to our findings in D-LDs. Recent findings from imaging
studies have shown that activation patterns in auditory cortex
were more related to the nature of the stimulus, whereas
patterns of activity in frontal area were more related to task
demands (Luo and others 2005). In frontal areas, task character-
istics, for example, same--different versus rhyming rather than
the stimulus processed (words vs. pseudowords or tones,
LoCasto and others 2004; Burton and others 2005), were found
to affect activation patterns. Burton and others (2005) sug-
gested that the task-related differences in activation are related
to differences in working memory load.
The behavioral impairments we measured are thus consistent
with each of the following 2 interpretations.
1. D-LDs’ deficits may be related to impaired transfer of
information between posterior, modality-specific stores
and the frontal areas carrying out specific working memory
operations (for review, see Curtis and D’Esposito 2003).
Such interpretation is consistent with the suggestion that
dyslexia is a ‘‘disconnection syndrome’’ in which communi-
cation is impaired between different cortical regions
(Paulesu and others 1996). In particular, Klingberg and
others (2000) have shown an abnormality of white matter
tracks in temporoparietal regions in dyslexics, which may
carry information from temporal to frontal regions.
2. Deficits may stem from impaired function of frontal areas.
Thus, information is properly transferred but is not properly
operated upon. Studies of executive functions in dyslexia
reported that subgroups of the dyslexic population are
impaired in tasks considered as reliable correlates of pre-
frontal function, such as the Wisconsin Card Sorting Test
(Helland and Asbjornsen 2000; Brosnan and others 2002).
Functional imaging studies also report abnormal patterns of
activation in prefrontal areas of dyslexic subjects (Backes
and others 2002; Shaywitz and others 2003). Furthermore,
the most notable change in activation pattern following an
effective auditory behavioral training (which improved
reading ability) was observed in prefrontal areas (Temple
and others 2003). The typical interpretation of these
findings is that dyslexics fail to utilize brain areas that are
normally involved in language processing. Language and
working memory are, however, hard to disentangle when
using verbal tasks. Although the amount of improvement in
language abilities in the study of Temple and others (2003)
was correlated with the magnitude of increase in activation
in the temporoparietal cortex, it could be that both are an
outcome of the prefrontal change.
The above ‘‘higher level’’ interpretations seem at odds with
previously reported findings of impaired activity of the auditory
Cerebral Cortex December 2006, V 16 N 12 1725
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

9. cortex. Thus, Nagarajan and others (1999) recorded deficits in
auditory cortical responses to tones presented with brief
intervals. Moreover, abnormal MMN responses to frequency or
phonemic oddballs were found among dyslexics (Kraus and
others 1996; Baldeweg and others 1999; Kujala and others 2003;
Maurer and others 2003; Renvall and Hari 2003; Lachmann and
others 2005). This apparent discrepancy probably results from
the following 2 reasons. First, the MMN response has multiple
generators, including frontal ones (e.g., Rinne and others 2000).
Impairment in the frontal generators seems to underlie the
deficits revealed in some of these studies, which reported an
abnormality in the later portion of the MMN response among
children at risk of dyslexia (e.g., Maurer and others 2003).
Second, other groups of dyslexics, who were not sampled in the
current study (which was focused on D-LDs), may have
a different, possibly ‘‘lower level’’ source of impairment. Pre-
liminary evidence that this may be the case was recently
reported by Banai and others (2005). They found that a sub-
group with a presumably early source of deficit (impaired
brainstem responses) had impaired MMNs. Interestingly, the
overall LD profile of this group was much milder (e.g., having
much better verbal memory) than that of D-LDs.
Analogies to Animal Studies
Consistent with the suggestion of a frontal source of impair-
ment, the pattern of D-LDs’ deficits with simple psychoacoustic
tasks resembles findings of studies characterizing dorsolateral
prefrontal cortex (DLPFC) functions in monkeys. These studies
report a similar dissociation between unimpaired perceptual
abilities and particularly poor working memory skills. Thus,
monkeys with bilateral lesions to DLPFC are impaired in a 3-
item task that requires them to remember which items they had
already chosen compared with monkeys with bilateral percep-
tual IT lesions (Petrides 2000). In this task, the same 3 elements
were presented in each of the 3 trial stages, though not in the
same positions. Monkeys had to touch only an item that they did
not choose in a previous stage of the trial. When reduced to a 2-
item task, which could be performed as a sequential S/D task,
monkeys with bilateral DLPFC lesion could reasonably perform
the task. We interpret this ability as reflecting a change-
detection strategy, which, unlike interstimulus comparisons,
can be subserved by low-level sensory circuitry (of inferotem-
poral (IT) cortex in this case). Interestingly, retention per se
seems to depend on the functioning of IT while comparisons
were contingent upon intact prefrontal cortex.
Parametric stimulus comparisons probably also involve pre-
frontal functions (Romo and Salinas 2003). Thus, neurons in
prefrontal cortex have been found active during the delay
period of a tactile frequency discrimination task similar to the
Hi/Lo task we used (Romo and others 1999; Romo and Salinas
2003), where this activity retained parameters of the 1st
stimulus. Though involvement of prefrontal cortex in auditory
working memory was not intensively studied, imaging studies in
humans (Belin and others 1998; Zatorre and others 1999) and
physiological (Romanski and Goldman-Rakic 2002) and anatom-
ical (Kaas and others 1999; Poremba and others 2003) findings
from nonhuman primates show that the prefrontal cortex is part
of the cortical network dedicated to the analysis of sound.
Taken together, these analogies suggest that the pattern of D-
LDs’ behavioral deficits is consistent with a mild dorsolateral
prefrontal deficit.
Notes
We thank Ehud Ahissar, Shabtai Barash, Eli Nelken, Udi Zohary, and
Ranulfo Romo for helpful comments on the manuscript. We thank the
Israel Science Foundation, Center of Excellence grant the Israeli
Institute for Psychobiology, and the Volkswagen Foundation for
supporting this study.
Address correspondence to Dr Merav Ahissar, Department of Psy-
chology, Hebrew University, Mt Scopus, Jerusalem 91905, Israel. Email:
msmerava@mscc.huji.ac.il.
References
Adlard A, Hazan V. 1998. Speech perception in children with specific
reading difficulties (dyslexia). Q J Exp Psychol A Hum Exp Psychol
51:153--177.
Ahissar E, Nagarajan S, Ahissar M, Protopapas A, Mahncke H, Merzenich
MM. 2001. Speech comprehension is correlated with temporal
response patterns recorded from auditory cortex. Proc Natl Acad
Sci USA 98:13367--13372.
Ahissar M, Protopapas A, Reid M, Merzenich MM. 2000. Auditory
processing parallels reading abilities in adults. Proc Natl Acad Sci
USA 97:6832--6837.
Amitay S, Ahissar M, Nelken I. 2002. Auditory processing deficits in
reading disabled adults. J Assoc Res Otolaryngol 3:302--320.
Amitay S, Ben-Yehudah G, Banai K, Ahissar M. 2002. Disabled readers
suffer from visual and auditory impairments but not from a specific
magnocellular deficit. Brain 125:2272--2285.
Backes W, Vuurman E, Wennekes R, Spronk P, Wuisman M, van
Engelshoven J, Jolles J. 2002. Atypical brain activation of reading
processes in children with developmental dyslexia. J Child Neurol
17:867--871.
Baddeley A. 1986. Working memory. Oxford: Clarendon.
Baldeweg T, Richardson A, Watkins S, Foale C, Gruzelier J. 1999.
Impaired auditory frequency discrimination in dyslexia detected
with mismatch evoked potentials. Ann Neurol 45:495--503.
Banai K, Ahissar M. 2004. Poor frequency discrimination probes
dyslexics with particularly impaired working memory. Audiol
Neurootol 9:328--340.
Banai K, Ahissar M. 2005. Psychoacoustics and working memory in
dyslexia. In: Syka J, Merzenich MM, editors. Plasticity and signal
representation in the auditory system. New York: Springer.
p 233--242.
Banai K, Nicol T, Zecker SG, Kraus N. 2005. Brainstem timing:
implications for cortical processing and literacy. J Neurosci
25:9850--9857.
Belin P, McAdams S, Smith B, Savel S, Thivard L, Samson S, Samson Y.
1998. The functional anatomy of sound intensity discrimination.
J Neurosci 18:6388--6394.
Ben Dror I, Shani M. 1998. Working memory sentence completion. Tel
Aviv, Israel: Mofet.
Ben-Yehudah G, Banai K, Ahissar M. 2004. Patterns of deficit in auditory
temporal processing among dyslexic adults. Neuroreport 15:627--631.
Ben-Yehudah G, Sackett E, Malchi-Ginzberg L, Ahissar M. 2001. Impaired
temporal contrast sensitivity in dyslexics is specific to retain-and-
compare paradigms. Brain 124:1381--1395.
Brosnan M, Demetre J, Hamill S, Robson K, Shepherd H, Cody G. 2002.
Executive functioning in adults and children with developmental
dyslexia. Neuropsychologia 40:2144--2155.
Burton MW, Locasto PC, Krebs-Noble D, Gullapalli RP. 2005. A
systematic investigation of the functional neuroanatomy of auditory
and visual phonological processing. Neuroimage 26:647--661.
Curtis CE, D’Esposito M. 2003. Persistent activity in the prefrontal
cortex during working memory. Trends Cogn Sci 7:415--423.
Daneman M, Carpenter PA. 1980. Individual differences in working
memory and reading. J Verb Learn Verb Behav 19:450--466.
De Weirdt W. 1988. Speech perception and frequency discrimination in
good and poor readers. Appl Psycholinguist 9:163--183.
Deutsch A, Bentin S. 1996. Attention factors mediating syntactic de-
ficiency in reading-disabled children. J Exp Child Psychol 63:386--415.
Fletcher JM, Shaywitz SE, Shankweiler DP, Katz L, Liberman IY, Stuebing
KK, Francis DJ, Fowler AE, Shaywitz BA. 1994. Cognitive profiles of
1726 Auditory Deficits in Dyslexia: Task or Stimulus d
Banai and Ahissar
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

10. reading disability: comparisons of discrepancy and low achievement
definitions. J Educ Psychol 86:6--23.
France SJ, Rosner BS, Hansen PC, Calvin C, Talcott JB, Richardson AJ,
Stein JF. 2002. Auditory frequency discrimination in adult develop-
mental dyslexics. Percept Psychophys 64:169--179.
Gathercole SE, Pickering SJ. 2000. Working memory deficits in children
with low achievements in the national curriculum at 7 years of age.
Br J Educ Psychol 70(Pt 2):177--194.
Hari R, Kiesila P. 1996. Deficit of temporal auditory processing in
dyslexic adults. Neurosci Lett 205:138--140.
Hari R, Saaskilahti A, Helenius P, Uutela K. 1999. Non-impaired auditory
phase locking in dyslexic adults. Neuroreport 10:2347--2348.
Heath SM, Hogben JH, Clark CD. 1999. Auditory temporal processing in
disabled readers with and without oral language delay. J Child
Psychol Psychiatry 40:637--647.
Helland T, Asbjornsen A. 2000. Executive functions in dyslexia. Neuro-
psychol Dev Cogn Sect C Child Neuropsychol 6:37--48.
Hochstein S, Ahissar M. 2002. View from the top: hierarchies and reverse
hierarchies in the visual system. Neuron 36:791--804.
Kaas JH, Hackett TA, Tramo MJ. 1999. Auditory processing in primate
cerebral cortex. Curr Opin Neurobiol 9:164--170.
Klingberg T, Hedehus M, Temple E, Salz T, Gabrieli JD, Moseley ME,
Poldrack RA. 2000. Microstructure of temporo-parietal white matter
as a basis for reading ability: evidence from diffusion tensor magnetic
resonance imaging. Neuron 25:493--500.
Kraus N, McGee TJ, Carrell TD, Zecker SG, Nicol TG, Koch DB. 1996.
Auditory neurophysiologic responses and discrimination deficits in
children with learning problems. Science 273:971--973.
Kujala T, Belitz S, Tervaniemi M, Naatanen R. 2003. Auditory sensory
memory disorder in dyslexic adults as indexed by the mismatch
negativity. Eur J Neurosci 17:1323--1327.
Lachmann T, Berti S, Kujala T, Schroger E. 2005. Diagnostic subgroups of
developmental dyslexia have different deficits in neural processing
of tones and phonemes. Int J Psychophysiol 56:105--120.
Leon MI, Shadlen MN. 2003. Representation of time by neurons in the
posterior parietal cortex of the macaque. Neuron 38:317--327.
Levitt H. 1971. Transformed up-down methods in psychoacoustics.
J Acoust Soc Am 49:467--477.
Liederman J, Kantrowitz L, Flannery K. 2005. Male vulnerability to
reading disability is not likely to be a myth: a call for new data. J Learn
Disabil 38:109--129.
LoCasto PC, Krebs-Noble D, Gullapalli RP, Burton MW. 2004. An fMRI
investigation of speech and tone segmentation. J Cogn Neurosci
16:1612--1624.
Luo H, Husain FT, Horwitz B, Poeppel D. 2005. Discrimination and
categorization of speech and non-speech sounds in an MEG delayed-
match-to-sample study. Neuroimage 28:59--71.
Lyon GR. 1995. Toward a definition of dyslexia. Ann Dyslexia 45:
3--27.
Maurer U, Bucher K, Brem S, Brandeis D. 2003. Altered responses to tone
and phoneme mismatch in kindergartners at familial dyslexia risk.
Neuroreport 14:2245--2250.
McAnally KI, Stein JF. 1996. Auditory temporal coding in dyslexia. Proc R
Soc Lond B Biol Sci 263:961--965.
Naatanen R, Tervaniemi M, Sussman E, Paavilainen P, Winkler I. 2001.
‘‘Primitive intelligence’’ in the auditory cortex. Trends Neurosci
24:283--288.
Nagarajan S, Mahncke H, Salz T, Tallal P, Roberts T, Merzenich MM. 1999.
Cortical auditory signal processing in poor readers. Proc Natl Acad
Sci USA 96:6483--6488.
Nicolson RI, Fawcett AJ, Dean P. 1995. Time estimation deficits in
developmental dyslexia: evidence of cerebellar involvement. Proc R
Soc Lond B Biol Sci 259:43--47.
Paulesu E, Frith U, Snowling M, Gallagher A, Morton J, Frackowiak RS,
Frith CD. 1996. Is developmental dyslexia a disconnection syn-
drome? Evidence from PET scanning. Brain 119(Pt 1):143--157.
Petrides M. 2000. Dissociable roles of mid-dorsolateral prefrontal and
anterior inferotemporal cortex in visual working memory. J Neurosci
20:7496--7503.
Poremba A, Saunders RC, Crane AM, Cook M, Sokoloff L, Mishkin M.
2003. Functional mapping of the primate auditory system. Science
299:568--572.
Putter-Katz H, Banai K, Ahissar M. 2005. Speech perception in noise
among learning disabled teenagers. In: Syka J, Merzenich MM,
editors. Proceedings of the symposium on plasticity of the central
auditory system and processing of complex acoustic signals. New
York: Springer. p 251--257.
Ramus F, Rosen S, Dakin SC, Day BL, Castellote JM, White S, Frith U. 2003.
Theories of developmental dyslexia: insights from a multiple case
study of dyslexic adults. Brain 126:841--865.
Raven J, Raven JC, Court JH. 2000. Manual for Raven’s progressive
matrices and vocabulary scales. Oxford: Oxford Psychologists Press.
Reed MA. 1989. Speech perception and the discrimination of brief
auditory cues in reading disabled children. J Exp Child Psychol
48:270--292.
Renvall H, Hari R. 2003. Diminished auditory mismatch fields in dyslexic
adults. Ann Neurol 53:551--557.
Rinne T, Alho K, Ilmoniemi RJ, Virtanen J, Naatanen R. 2000. Separate
time behaviors of the temporal and frontal mismatch negativity
sources. Neuroimage 12:14--19.
Romanski LM, Goldman-Rakic PS. 2002. An auditory domain in primate
prefrontal cortex. Nat Neurosci 5:15--16.
Romo R, Brody CD, Hernandez A, Lemus L. 1999. Neuronal correlates of
parametric working memory in the prefrontal cortex. Nature
399:470--473.
Romo R, Salinas E. 2003. Flutter discrimination: neural codes, percep-
tion, memory and decision making. Nat Rev Neurosci 4:203--218.
Rosen S. 1999. A problem with auditory processing? Curr Biol 9:
R698--R700.
Share DL, Silva PA. 2003. Gender bias in IQ-discrepancy and post-
discrepancy definitions of reading disability. J Learn Disabil 36:4--14.
Shaywitz S. 1998. Dyslexia. N Engl J Med 338:307--312.
Shaywitz SE, Shaywitz BA, Fletcher JM, Escobar MD. 1990. Prevalence of
reading disability in boys and girls. Results of the Connecticut
Longitudinal Study. J Am Med Assoc 264:998--1002.
Shaywitz SE, Shaywitz BA, Fulbright RK, Skudlarski P, Mencl WE,
Constable RT, Pugh KR, Holahan JM, Marchione KE, Fletcher JM,
Lyon GR, Gore JC. 2003. Neural systems for compensation and
persistence: young adult outcome of childhood reading disability.
Biol Psychiatry 54:25--33.
Siegel LS. 1992. An evaluation of the discrepancy definition of dyslexia.
J Learn Disabil 25:618--629.
Siegel LS, Ryan EB. 1989. Subtypes of developmental dyslexia: the
influence of definitional variables. Read Writing 1:257--287.
Snow CE, Burns MS, Griffin P, editors. 1998. Preventing reading diffi-
culties in young children. Washington, DC: National Academic Press.
Snowling MJ. 2000. Dyslexia. Oxford: Blackwell.
Swanson HL. 1993. Working memory in learning disability subgroups.
J Exp Child Psychol 56:87--114.
Talcott JB, Witton C, McClean M, Hansen PC, Rees A, Green GG, Stein JF.
1999. Can sensitivity to auditory frequency modulation predict chil-
dren’s phonological and reading skills? Neuroreport 10:2045--2050.
Tallal P. 1980. Auditory temporal perception, phonics, and reading
disabilities in children. Brain Lang 9:182--198.
Tallal P, Miller S, Fitch RH. 1993. Neurobiological basis of speech: a case
for the preeminence of temporal processing. Ann N Y Acad Sci
682:27--47.
Temple E, Deutsch GK, Poldrack RA, Miller SL, Tallal P, Merzenich MM,
Gabrieli JD. 2003. Neural deficits in children with dyslexia amelio-
rated by behavioral remediation: evidence from functional MRI. Proc
Natl Acad Sci USA 100:2860--2865.
Tramo MJ, Shah GD, Braida LD. 2002. Functional role of auditory cortex
in frequency processing and pitch perception. J Neurophysiol
87:122--139.
Vargo FE, Grosser GS, Spafford CS. 1995. Digit span and other WISC-R
scores in the diagnosis of dyslexia in children. Percept Mot Skills
80:1219--1229.
Vellutino FR, Fletcher JM, Snowling MJ, Scanlon DM. 2004. Specific
reading disability (dyslexia): what have we learned in the past four
decades? J Child Psychol Psychiatry 45:2--40.
Cerebral Cortex December 2006, V 16 N 12 1727
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom

11. Waber DP, Wolff PH, Forbes PW, Weiler MD. 2000. Rapid automatized
naming in children referred for evaluation of heterogeneous learning
problems: how specific are naming speed deficits to reading
disability? Neuropsychol Dev Cogn Sect C Child Neuropsychol
6:251--261.
Wechsler D. 1998. Wechsler intelligence scale for children (R-
95,1998 Israeli edition)—manual. San Antonio, TX: The Psycholog-
ical Corporation.
Witton C, Talcott JB, Hansen PC, Richardson AJ, Griffiths TD, Rees A,
Stein JF, Green GG. 1998. Sensitivity to dynamic auditory and visual
stimuli predicts nonword reading ability in both dyslexic and normal
readers. Curr Biol 8:791--797.
Woerner C, Overstreet K, editors. 1999. Wechsler abbreviated scale of
intelligence (WASI). San Antonio, TX: The Psychological Corporation.
Wright BA, Bowen RW, Zecker SG. 2000. Nonlinguistic perceptual
deficits associated with reading and language disorders. Curr Opin
Neurobiol 10:482--486.
Zatorre RJ, Mondor TA, Evans AC. 1999. Auditory attention to space
and frequency activates similar cerebral systems. Neuroimage 10:
544--554.
1728 Auditory Deficits in Dyslexia: Task or Stimulus d
Banai and Ahissar
byguestonJune29,2011cercor.oxfordjournals.orgDownloadedfrom