Combined auditory and articulatory training improves
phonological deficit in children with dyslexia
Barbara Joly-Pottuz, Ph.D.
Melina Mercier, Ph.D.
Aurelie Leynaud, M.A.
Michel Habib*, M.D., Ph.D.
Department of Paediatric Neurology
University Hospital and Institut de Neuroscience Cognitive de la Méditerranée, CNRS,
Submitted to Neuropsychological Rehabilitation
August 23, 2006
Revised version (2), May 2007
*Correspondence to last author at Service de Neurologie Pédiatrique, CHU Timone,
13385 Marseille Cedex 5, France
fax # : (33) 4 91 46 07 34
E-mail : firstname.lastname@example.org
Keywords : phonology, training, articulation, dyslexia, cerebellum
Running title : articulatory and auditory training in dyslexics
19 children with dyslexia aged 7;2 to 10;9 were selected from a clinical sample and
tested using a large neuropsychological battery in order to specify the severity and
subtype of dyslexia as well as the presence of comorbid conditions. Thereafter, they
received a standardized training of 6 weeks of daily auditory exercises aimed at
reinforcing explicit and implicit phonological awareness. 10 participants also received
a specific training of the sensory-motor aspects of articulatory production of
individual phonemes during the first 3 weeks of auditory training, whereas the
remaining received the same during the last 3 weeks of training. Repetition,
phonological awareness, reading and spelling were assessed before the first session,
between the two sessions and after the second session. Results confirm the overall
efficiency of intensive phonological training, even with exclusively auditory material.
The main outcome of this study is a significant improvement of phonology and non-
word reading specifically during the periods where the two methods were associated,
suggesting a significant contribution of articulatory training to the observed
improvement. Finally performance to a motor tapping task proved to be one of the
best predictors of training efficiency while comorbid coordination or attention deficit
did not interfere. Results are interpreted in reference to current theories about
mechanisms underlying dyslexia.
This work was made possible thanks to an “ACI Cognitique – 1999-2000” funding from the
French Ministry of Scientific Research and “PHRC-APHM 2001” grant. Barbara Joly-Pottuz
doctoral thesis was partly supported by ADREN, RESODYS and APHM.
The authors wish to thank three reviewers for their valuable comments.
Remediating dyslexia has become a challenge for both neuroscientists and clinicians
during these last ten years for several reasons. First, dyslexia, the commonest form of
learning disability, is a widespread condition affecting about 5-7 % of school-age
children with major impact on their academic and ultimate social achievement (Habib,
2000; Démonet et al., 2004). Moreover, scientifically proven efficient training is often
taken as evidence in favour of the mechanism postulated to be causal to the deficit, thus
contributing to expanding knowledge about learning disorders and their determinants.
In this connection, the mechanism most often invoked at the origin of dyslexia is that of
poor phonological coding due to deficient phonological representation (Swan &
Goswami, 1997; Snowling, 2001; Ramus et al., 2003c). This phonological deficit usually
shows up as a reduced ability to discriminate and manipulate speech-sounds in spoken
words, so-called phonological awareness. Such deficit, which can be traced back to early
childhood, eventually results in a profound incapacity to learn associations between
written and oral codes (grapheme-to-phoneme mapping), typically an important stage
toward expert reading (Stanovitch, 1988; Rack et al., 1992). Accordingly, studies
reporting successful phonological training in dyslexia are abundant and their results
largely consistent across studies, to such an extent that this constitutes the universal
basis for dyslexia therapy (Ehri et al., 2001; Torgesen et al., 2001; Wise et al, 1999;
Vellutino et al, 2004; McCandliss et al., 2003; Ortiz-Gonzalez et al., 2002) as well as for
interventions carried out in at risk children to prevent subsequent reading difficulties
(Foorman et al., 1998; Alexander & Slinger-Constant, 2004; Elbro & Petersen, 2004). In
addition to strictly phonological training, there is also recent evidence that training
children to simple auditory discrimination of phonemes may result in significant
improvement on phonological as well as on reading tasks (Moore et al., 2005). Finally,
such improvement in reading-related tasks has been related to changes in brain
activation patterns, as demonstrated with various brain imaging techniques (Simos et
al., 2002, Aylward et al., 2003, Temple et al., 2003; Kujala et al., 2001, Shaywitz et al.,
However, all these results suffer from two important weaknesses which may hinder
their validity and reproducibility : first, recruitments are generally made from classroom
populations, and thus probably include children with low reading ability who would
otherwise not meet strict criteria for dyslexia, also likely gathering indistinctly dyslexics
of various subtypes. Another– and probably more problematic– concern with these
studies is that, for the sake of efficacy, researchers have generally used remediation tools
that were already available at the moment of the study, usually covering wide spectra of
underlying mechanisms of presumed impairment instead of focusing on one or few
processes. As a consequence, all these studies, though otherwise valuable in showing
the potential for children to improve, are barely suitable for a fine-grained analysis of
the actual mechanisms underlying improvement. For instance, most of these tools are
claimed to provide ‘phonic’ interventions, a term which encompasses a large range of
processes including phonology (the sound structure of speech), phoneme
discrimination, and acquisition of connections between these and letters (phonological
One of these is the method proposed by Tallal and co-workers, known as
“FastForword®” (FFW: Tallal, et al., 1996; Merzenich et al., 1996). On the basis of
observations suggesting that the dyslexic brain would suffer from a specific inability to
process brief and rapidly presented stimuli (the “temporal-processing deficit theory of
dyslexia”, Tallal, 1980), these authors have implemented a series of video-games for
which the auditory component includes temporally modified speech in order to
facilitate and progressively restore brain temporal processes. In this and other similar
complex and multifaceted, interactive audio-video games (Kujala et al., 2001; Agnew et
al., 2004; Hayes et al., 2003b; Magnan et al., 2004), it could be difficult to separate the
effects of auditory and visual linguistic stimulation, as well as to disentangle attentional
from motivational factors. For example, in a study summarizing five separate trials of
FFW training program, Gillam et al. (2001b) reached the following conclusion : “…It is
possible that similar improvements … may be obtained from a variety of interventions that are
presented on an intensive schedule, that focus the child’s auditory and visual attention, that
present multiple trials…, and that reward progress”.
In a series of preliminary studies, and in an effort to avoid such important biases, Habib
et al. (1999, 2002) devised a specific protocol of phonic intervention exclusively
comprising auditory exercises tackling the phonological system, in order to reduce the
possible intervention of confounding factors. Children received daily exercises where
they only had to indicate manually, by pointing to one of three labels, which one out of
3 words or non-words successively heard into headphones was phonetically different
from the other two. In addition, in order to limit interference with other (including
spontaneous) factors of recovery, the training sessions were compacted over a relatively
short period (5 to 6 weeks). Note that with this protocol, which was initially elaborated
to test Tallal’s temporal deficit theory by comparing intensive training with and without
temporally modified speech, authors only found a modest advantage of training
children with slowed speech (Habib et al., 1999). In contrast, they showed that,
notwithstanding the actual speech form (with or without temporal modification of the
signal), children markedly improved their performance on phonological and spelling
tasks, with, however, only partial generalization to reading (Habib et al., 2002).
The first aim of the present study was then to confirm this preliminary observation in
showing that children with dyslexia do benefit from such daily exercises, exclusively
involving auditory spoken materials, during a relatively short period, and even under
unmodified form. Since such training is virtually devoid of any visual component, we
anticipate that the resulting gain in reading will concern “phonological reading”
(reading non-words) rather than “sight reading” (reading irregular and exception
Another objective for the present study derives from previous evidence that articulatory
exercises may be of some benefit to children with dyslexia, an issue first raised by
Alexander et al. (1991). These authors gave a series of preliminary tests to 10 children
with dyslexia, and found a marked phonological deficit, as expected. Then, they trained
the children individually with a program designed to enhance their awareness of
articulatory movements during speech production. This program sought to integrate
tactile and somatosensory information from articulators, visual information provided by
a mirror placed in front of the child in such a way to allow him/her to see his/her own
mouth during speech production, and finally auditory feedback resulting from his/her
own sound production. This training program included tutoring participants with
schematic drawings representing the form of vocal apparatus for each trained phoneme.
Although involving a limited number of participants, this study yielded encouraging
results. However, subsequent work by Wise et al. (1999), on a much larger group (122
subjects), failed to find significant differences in phonological reading (non-word
decoding) between those poor readers similarly trained for articulatory sensory-motor
integration, and those only trained with phonological manipulation during the same
period.. Indirect support in favour of the utility of such articulatory training comes from
the proposal by Nicolson and Fawcett (1999) of a cerebellar origin of dyslexia. More
specifically, these authors suggested that two frequent aspects of dyslexic children
deficits, the incidence of which has been underestimated, i.e. motor coordination and
automation deficits, may be ascribed to cerebellar dysfunction. This possibility is
supported by clinical (Stoodley et al., 2005), anatomical (Finch et al., 2001) and brain
imaging (Nicolson et al., 1999) evidence. The proponents of this theory reasoned that
while reading and spelling ability would be mainly affected by the deficit in
automaticity, the phonological deficit would emerge through poor articulatory ability.
However, this motor theory is currently severely criticized by several groups (Raberger
& Wimmer, 2003, Ramus et al., 2003a,b), suspecting that the observed occurrence of
cerebellar symptoms would be actually an artefact resulting from the comorbidity of
dyslexia with other conditions such as dyspraxia (or coordination disorder) and
In the present work, we therefore sought to evaluate the potential benefit of training
sensory-motor aspects of articulatory function in addition to auditory phonological
training, and the clinical usefulness of combining auditory and articulatory stimulation,
in a group of children with dyslexia in whom possible comorbid coordination or
attentional disorders were assessed. Thus, careful selection of participants, together with
optimized separation of factors of potential improvement, were expected to maximize
the reliability and thus practical validity of the use of such methods in clinical settings.
Material and methods
Nineteen children, aged 7;2 to 10;9, were selected on the basis of (a) a clinical diagnosis
of dyslexia (see below), (b) current enrolment in speech-language therapy (usually 1 or 2
weekly sessions, in accordance with usual practice in France, , and (c) acceptance to
participate in a research program.
The diagnosis of dyslexia was made by the speech-language therapist, and confirmed by
the research team, based on (1) a report of significant difficulty in beginning to read
(mostly absence of mastering of grapheme-phoneme mapping after one year of
conventional teaching); (2) the reading difficulty being non explainable by lack of
intelligence or insufficient schooling; and (3) performance on at least one reading task
more than one s.d. below that expected for age.
Once thus selected, all children received an evaluation of non-verbal IQ with Raven’s
Progressive Matrices and several subtests (similarities, block design, picture completion
and digit span) of the Wechsler Intelligence Scale for Children, Third Edition.
Children were excluded if they scored below 25th percentile at the PM 47 and/or below
8/19 on the WISC similarities subtest.
For increasing the homogeneity of the group, we chose to exclude children with
expressive and/or receptive oral language problems at the time of inclusion, as
evidenced by normal or near-normal performance on classical batteries used by speech-
language therapists1 .Especially , as expected with a main diagnosis of dyslexia, none of
the 19 participants had apparent articulatory difficulty (which does not rule out more
subtle, subclinical deviances; see Lalain et al., 2003).
Dyslexia severity was measured using the classical French “test de l’Alouette”
(Lefavrais et al., 1965), providing a reading age in months. Besides tasks included in
selection assessment, children were also assessed for usual comorbidities of dyslexia :
mainly attention deficit and coordination deficit (here not considered exclusion criteria),
as well as general executive dysfunction.
Oral fluency (Controlled Oral Word Association test) was carried out under two forms
(Chevrie-Muller et al., 1997) : phonetic and semantic, asking children to provide in one
minute as many words as possible starting with letter P and letter F (phonetic) and as
many words as possible related to a given category (sport, professions, vacation).
Response inhibition was tested with a modified version of the Stroop task (“animals”
Stroop task, Jacot-Descombes & Assal (1986) which consists in 3 printed plates
representing either a series of 50 drawings of 5 animals horizontally arranged in semi-
1 Note that, in the French Health system, speech-language therapists are traditionally in charge of
diagnosis and remediation of reading/spelling pathologies, which is not the case in other
random order (plate 1), or printed word of the same animals (plate 2), or finally, a
superposition of drawings with the name of another animal (plate 3). Time needed for
naming all items on plate 3 weighed against that needed to name items on plate 1 serves
to calculate an “interference index”, reflecting the individual’s capacity to inhibit
spontaneous tendency to read the superposed word. Plate 1 also served as a measure of
rapid naming. Z-scores were calculated by reference to age-matched controls.
Children were also tested specifically as to the possible diagnosis of comorbid Attention
Deficit -Hyperactivity Disorder (ADHD), according to results to Conners’ questionnaire
for parents (Conners, 1989), which also served to rule out other psychiatric conditions
such as anxiety traits or behavioural deviances which could interfere with children’s
compliance to treatment. Only two participants scored 15/30 or more on the Conners’
questionnaire, thus possibly diagnosed as ADHD.
Visuo-spatial abilities were tested using the classical Rey’s figure copy (Rey, 1959), as
well as part of the Signoret’s “Batterie 84” for visuo-spatial memory (Signoret, 1991).
Short-term memory was selectively explored by classical span tasks, in the auditory-
verbal and visuo-spatial modalities. A coordination deficit was suspected when parents
report included instances of delayed motor development, and indirectly measured by
the “Stambak test”, a classical tapping task involving the reproduction of rhythmic
successions of sounds given under patterns of increasing complexity (Stambak, 1951).
Finally, although some participants may still have some subtle motor difficulties, none
of them qualified for the main diagnosis of dyspraxia or developmental coordination
disorder (Visser, 2003). Nonetheless, a history of slightly delayed motor milestones
and/or evidence of dysgraphia, diagnosed both from parents report and direct
observation of graphomotor productions, was present in six participants. Copy of the
Rey figure was also often awkward or slightly disorganised.
In sum, all 19 participants to this research had a classical picture of moderate to severe
dyslexia, without persistent production or reception speech/language pathology, but
with usual comorbidity, mainly moderate attentional and/or motor coordination and
slight spatial difficulties.
Each of the following tasks was administered 3 times at E1, E2 and E3 (see below,
Children were required to repeat a series of simple (CV , e.g.: /ba, da, ga/), complex
(CCV, e.g. /gra, dra/) syllables, semivowels (e.g. : /oui/) and nasal vowels (e.g; : /ã,
õ/), simple and complex words. Productions are rated according to the proximity or
distance to each specific target, as judged by a trained speech therapist. A score is given
for each list, and a global score calculated as the average score of all lists. These tasks are
supposed to explore not only articulatory precision and agility but also, to a lesser
degree, auditory short-term memory. This latter aspect was more specifically targeted
by a separate sentence (4 items) and non-word repetition task (two lists of 20 CV or
CCV, 1 to 5-syllable pseudo-words of increasing difficulty).
Phonological awareness (P.A.)
P.A. tasks were taken from a frequently used battery in French (BELEC, Mousty et al.,
1994), exploring segmenting, discrimination and short-term memory abilities. All tasks
were played by a tape-recorder to insure homogeneity of testing conditions. These
included tasks looking at various levels of P.A. : syllable inversion (10 bisyllabic CVCV
pseudowords, e.g.: /radi/ —>/dira/)) phoneme inversion (10 CV and 10 VC
monosyllabic items; e.g: /bi/—>/ib/ and /ol/ —>/lo/), syllable deletion (16 items, e.g.
: /fango/ —> /go/), simple (10) and complex (10) phoneme deletion (e.g.:
/gal/—>/al/; /grõ/—>/rõ/), auditory acronyms (16 word pairs : form a new word
with the initial sound of each of the two words).
In addition, a rhyme judgment task of the odd-one-out type, similar to those used in the
training exercises (see below) was used : it comprised 10 bisyllabic word triplets, the
subject being asked to indicate the 2 rhyming ones (ex: ballon-cochon-robot = rhyming
target : /õ/).
Finally, a “phonological awareness score”was calculated by averaging performance
from these different tasks.
Due to limited controlled materials being available in French, reading tasks were
different for younger (grades 1&2, i.e. 6-8 y-old) and older (grades 3-4, i.e. 8-10 y-old)
children. For grades 1-2, we used a locally standardized battery (Hénin & Dulac, 1985)
comprising an extensive, progressive exploration of all the difficulties of the French
orthographic system (simple graphemes – syllables – multiple consonants, e.g.: [olp] or
[spli] - complex graphemes, e.g; : [ain] read /C/ or [ien] read /JC/ - ambiguous forms,
e.g. [payé] read /peJe/) and, finally, a simple word reading task (24 items). There was
no irregular words list for this younger group, due to their relative rarity in French, so
that “sight” reading, at this age, could only be tested by reading regular words. For
grades 3-4, children were asked to read aloud 3 lists (pseudowords : n=20; regular
words, n=10; irregular words n=10) taken from the French L2MA battery (Chevrie-
Muller et al., 1997).
All recorded performances were transformed into a percentage of correct responses in
order to equate measures across children of different age ranges. A “phonological
decoding score” was calculated from the performance to all non-word items, and a
“word reading score” from performance to all lexical items from the two age groups.
For the same reasons as above, spelling tasks were different for grade 1-2 (transcription
to dictation of 20 words and 4 simple sentences), and grades 3-4 (transcription of a short
story « Le Petit Poucet » (Hénin & Dulac, 1985).
For all these spelling tasks, errors were classified according to the suspected underlying
mechanism (phonological, grammatical, or usage errors). However, for sake of
simplicity, we only used a global spelling score in terms of error rate.
A separate non-word spelling task was also given, consisting of eight simple (mainly
CVCV) and 15 complex (CCV) pseudo-words
INSERT HERE FIGURE 1
Experimental design (figure 1)
Our experimental protocol followed an A-B vs B-A design, where each group is its own
control in order to avoid ‘carry-over effects’ due to the risk that the treatment used in
the first period affects the outcome of the second period. According to this protocol, 10
children were randomly assigned to group 1 (G1), receiving during 3 weeks (period A)
both auditory-phonological training and articulatory training, then only auditory
training during the next 3 weeks (period B), whereas the remaining 9 children, assigned
to group 2 (G2), received the same total amount of training, but in the opposite order,
i.e. in period A, only auditory training, followed by auditory + articulatory training, so
as to counterbalance a possible order effect. Note that this design has the advantage that
there is no control group which could have its own effect (see Wise et al., 1999), but the
inconvenience that periods A and B are not equivalent in terms of overall amount of
therapy during the 3-week period.
Auditory phonological training
The method involves daily listening by children with dyslexia of series of exercises,
recorded on audio-CDs and mainly containing triplets of words the patient had to
compare phonologically to disclose similarity between 2 of the 3 stimuli, either at the
beginning, the end, or at the middle of words (“odd-one-out” task). The only indication
the children had was about the locus of the expected change in the word and whether
they must focus on syllables (weeks 1-3) or sounds (weeks 4-6), each specific instruction
being recorded and played before each exercise. Each daily session, corresponding to
one track of the CD, lasted 15-25 mn, and was administered under the control of one
parent, asked to report the child’s response on a pre-printed response sheet. Participants
and their parents were informed that the study will involve two parts, evaluation and
training, and parents were informed that they will have to participate actively in the
latter. Indeed, such participation of parents was crucial to the project. We particularly
insisted on the need to avoid any feedback, since parents could have the spontaneous
tendency to give the correct answer. Moreover, there was no visual component, and no
oral production was needed since participants only had to point to a specific label. Once
a week, information on the unfolding of the training procedure was obtained either by a
visit at home or by telephone, and the response sheets were asked back to measure the
level of compliance and make sure that the treatment is correctly carried out.
The training program has been designed under the principle of non-adaptive
progression from simpler to harder phonological tasks. During the first 3 weeks,
exercises only involve syllable segmentation, with increasing difficulty due to the
closeness of the distractor to the target words. For example, at week 3, children have to
discriminate between syllables containing phonemes of the same voiced-unvoiced pair
to “find the 2 words with the same syllable in the middle”
(“député/carburant/débutant”), with the additional difficulty of a common syllable
between the distractor and one target word (/de/ in first syllable). During weeks 3 to 6,
solving tasks requires detecting the common sound in two words, so that there is a
progression from syllabic to phonemic segmentation requirements. At week 4,
distractors are unrelated phonologically, whereas at week 6 they are (example :
“palais/panda/bandeau” “find the two words beginning with the same sound”, where
the rest of the syllable and middle phoneme are common between distractor and one of
the target words). Through such manipulation of target-distractor closeness, of course
unknown to participants, we wished to tackle not only explicit (metaphonological) but
also implicit phonology (Boada & Pennington, 2006).
Articulatory training involved the combination of two complementary approaches, both
used in the same session, one aiming at a reinforcement of the participants’ articulatory
awareness, the other one using possibilities offered by a computerized training tool. It
was carried out 3 times a week, in a one-to-one setting, each session being monitored by
a speech therapist or a neuropsychologist.
a) Articulatory awareness exercises.
This part of the protocol is based on evidence in dyslexics of low awareness of their
articulatory system (Griffiths and Frith, 2002), leading to the idea that even in the
absence of objective motor or articulation problem, the sensory-motor speech system
may usefully be reinforced by specifically becoming aware of each individual phoneme
with the support of proprioceptive, tactile and visual information.
Exercises were derived from those described in Alexander et al. (1991), putting
emphasis on phonetic oppositions between voiced (/b/, /d/, /g/) and voiceless stop
consonants (/p/, /t/, /k/).
The first three sessions were devoted respectively to : p/b ; t/d and k/g contrasts, then
all the pairs were worked on again during the subsequent sessions. Each session
included the same exercises in the same order :
1- Practice of voiceless phoneme (e.g; /p/) : The child was asked to produce the
phoneme in front of a mirror in order for him/her to become aware of the
position of lips during the aural production. Likewise, he /she will be taught
how to recognize with finger palpation of his/her neck the sensation associated
with unvoiced production.
2- Similar exercises with the voiced phoneme (e.g.: /b/).
3-Presentation and explanation of the anatomical sketch associated to this specific
pair of phonemes : the child was trained to associate each stop consonant to the specific
corresponding scheme of speech organs, a series of drawings inspired from a classical
French aphasiology treatise (Lanteri, 1995), each representing the position of lips,
tongue, teeth and palate during the production of a specific phoneme. Presence or
absence of voicing was figured by a coloured horizontal stroke, a green, straight one for
voiceless, and a red, wavy one for voiced consonants.
4- Matching to sketch : after the examiner had pronounced one given word, the
child was asked to extract the initial phoneme of the word and to point to the
The specific training words have been chosen as to be « paronymes », which means that
they are bisyllabic only differing by one phoneme situated in one of the syllables
(palais/balai; touche/douche) and varying from one session to another. Each list
contains six word-pairs.
b) Phoneme production with computerized visual feedback
This part of the training method takes advantage of facilities included in IBM Speech-
viewer ® III (SPV3) program, a specialized software for professionals in speech
pathology. Children had to pronounce a chosen phoneme through a microphone, while
the program analysed the acoustic quality of the production. The utterance is then
transformed into a visual index, for example a coloured part of the picture which
changes from green to red, or a slope which is being climbed by a character. The
following exercises have been chosen for practice of voicing : « Voice Presence », which
develops an awareness of voicing (a clown changes colour, a man’s lips move) ,« Voice
Onset » supposed to increase the awareness of voicing onset and control over voicing,
and « Voice Timing », which improves coordination of respiration and voicing. For each
exercise, the participant was asked to utter specific phonemes or groups of phonemes as
accurately as possible, with the support of the visual index provided on the screen,
specific to each game.
During the first 3 sessions, these games are being used in succession for practice of
voiced/voiceless contrast with pairs of stop consonants, with in addition for each
session one pair of constrictive consonants with similar voicing opposition (f/v, ch/j,
In addition, other parts of the program were used to complete the above effects :
« Phoneme Accuracy », which improves the accuracy of phoneme production (using a
fun game format) ; « Multi-phoneme Chains » developing skills in pronouncing a
sequence of different phonemes, and « Two-Phoneme Contrast » and « Four-Phoneme
Contrast », both designed to improve accuracy in contrasting phonemes.
For instance, for the two-phoneme contrast, the aim of the game is to lead a vehicle
through a series of obstacles by pronouncing the selected phoneme.
A different contrast is associated to each specific vehicle (picture of a car : f/v; picture
of a bike : ch/j, picture of a Jeep : s/z)
The speed of movement of the vehicle and the threshold of recognition of each phoneme
can be modified so as to provide an adaptive control to each child’s performance.
Finally, each participant received :
daily 20-25-mn sessions of auditory phonological training, made up of
auditory exercises, without feedback and non adaptive
35-40-mn sessions of articulatory training, combining both approaches
(articulatory awareness and Speech-Viewer games), 3 times a week, with
explicit feedback and adaptive progression of the game content.
Overall, pre-testing began in October, training began in January and extended over a
period of 3 months, all participants having received their 6 weeks of training by the end
of April. On average, each child had received 14 hours of auditory training and 12 hours
of articulatory exercises.
The 3 evaluation sessions (E1, E2, E3) respectively took place during the week preceding
the beginning of training, after the first 3 weeks of training and after completion of the 6
weeks of training.
A 2-way repeated-measure ANOVA, considering time as the within subject factor, was
used to test for different effects of session by group, i.e. one group’s improvement
follows a temporal course which is significantly different from the other group. Our a-
priori hypotheses are (1) : there will be a general effect of time, meaning overall
improvement for both groups; (2) improvement will involve primarily tasks dependent
on phonological processes; (3) for these tasks, the two periods of combined auditory and
articulatory treatment (period A for group 1 and period B for group 2) should result in
better improvement. Note that with such an experimental design, one would not expect
to find a significant group x time interaction, suggestive of an order effect. More
probably, there will be an absence of interaction, reflecting similar cumulative
improvement during the first and the second periods of training. Therefore, post-hoc
pair-wise comparisons will be necessary to measure the direction and amplitude of each
Finally, we calculated correlation coefficients for several measures at E1 to test for their
predictive value on training efficacy, measured as an improvement index (E3-
E1/E3+E1). In line with the general context of this research, we more specially focused
on tasks having a motor component (repetition and motor tapping tasks) and tasks
exploring rapid word recognition, rapid naming, and response inhibition. Improvement
index was calculated for phonology, reading, spelling, and repetition. Correlations at E1
were also computed between all these variables.
Parents were met as a group and individually and given detailed explanation of the
training methods as well as about their role in this research. All children and parents
were informed of the aims and principles of the study, and both parents signed
informed consent. The study was authorized by the local Committee of Ethics
INSERT HERE TABLES 1&2
Comparisons between the characteristics of the two groups are presented on tables 1
and 2. The two groups did not differ significantly in chronological age, reading age, and
overall severity of dyslexia (table 1). There was a non-significant tendency for group 2
toward a better verbal (WISC similarities) and non verbal (Raven’s PM47) performance
than group 1. The main features are representative of a population of dyslexics of
moderate severity, with 10-20 months difference between reading and chronological
ages, and deficient phonetic fluency. Reading delay was larger in group 2, but this
was not significant and did not affect intra-group comparisons (see below), as was the
case for the marginal difference in non-verbal intelligence.
Table 2 summarizes participants’ performance on some specific neuropsychological
tasks. Rapid automatized naming as well as rapid word recognition were both
moderately, but significantly, below expectations, without significant difference
between the two groups. Likewise, response inhibition as measured by the naming time
on the Stroop condition tended to be poorer than normal, without significant difference
between groups. As concerns the rhythm reproduction task, performances also tended
to be poorer than in controls of the same age, but here again remained within the low
average range. Overall, the two groups did not differ as regards motor coordination
(Stambak task), response inhibition (Stroop-like task) or hyperactivity rating using the
Conners’ score. There was no correlation between the Stroop task and any other task.
INSERT HERE TABLES 3&4
The main performances on phonological tasks are reported on table 3 . Apart from
syllabic segmentation, all measures were lower than expected for age, although often
marginally, except for repetition tasks where scores in percentiles were more clearly in
the pathological range. . It is not surprising to find individual phonological scores falling
within the normal range, since most of them had already received phonological training
prior to this study. There were no significant differences between groups 1 and 2.
Table 4 and figure 2 summarize the mean and standard deviations obtained at E1, E2
and E3 for composite scores calculated from repetition tasks, spelling tasks,
phonological awareness and reading tasks. Simple effects of time (session) were largely
significant for all variables, suggesting overall improvement for all domains explored.
No group effect was obtained on either variable, and no group x session interaction was
present. In particular, despite a larger reading delay in group 1, the two groups did not
differ in overall reading improvement.
Statistics relevant to E1 vs E2 and E2 vs E3 comparisons lead us to consider three
different patterns of results :
typical additive effect : for phonological awareness and phonological decoding
scores, the pattern of evolution over time during the training period was strongly
suggestive of an additive effect of articulatory on phonological training, with significant
improvement during the first training period for G1 and the second training period for
G2. As shown in figure 2a and b, although improvement was progressive from E1 to E3
for both groups, it seems to plateau between E2 and E3 for group 1 and between E1 and
E2 for group 2, i.e. during the “only phonology” periods. A similar tendency was
observed for word and text spelling tasks.
INSERT HERE FIGURE 2
for repetition and non-word spelling tasks , the only significant improvement
occurred between E1 and E2, for both groups, suggesting that articulatory training has
had no cumulative effect on phonology training : training phonology, with or without
articulatory awareness component, significantly improved repetition and non-word
spelling during the first three weeks, but this tendency was short-living and vanished
later on. Alternatively, this could reflect a ceiling effect, especially for the repetition task
where the performance on E2 is already near maximal.
Finally, a third pattern, here represented by regular and, for the older
participants, irregular word reading tasks (thus exploring the lexical route of reading), is
that of delayed improvement (only significant between E2 and E3), without difference
between the two groups, meaning that reading words from sight takes longer to benefit
from phonological exercises, whether or not accompanied with articulatory training
(figure 2c)2. This pattern seems to reveal potential long term effect of combined
treatment on reading. However, in the absence of follow-up measures, it is hazardous to
speculate about possible subsequent benefit.
2 Note that progression between E1-E2 and E2-E3 may be significantly different even
though the slopes do not seem different.
Figure 3 reports correlation analyses carried out between neuropsychological variables
and some pre-test measures on the one hand, and between these same variables and
pre/post-training improvement of composite scores on the other hand.
The most significant results are those obtained between pre-test repetition scores, and
phonological efficiency, as measured by pre-test phonological awareness scores (r=0.535
p=0.0182). In other terms, there is statistical evidence of a link between production
efficiency and phonological awareness, but one can hardly speculate further on the exact
nature of this link.
The last issue to be considered here is that of the predicting value of some pre-test
measures on training efficacy, which we chose to address through simple correlation
analyses : one of the best predictors of pre-post training improvement was performance
on the “Stambak task”, which significantly predicted reading (r=0.566; p=0.0115) and
non-word spelling (r=0.656, p=0.0023) improvement. Rapid naming at E1 also predicted
improvement on phonological awareness tasks (r=0.643; p=0.003), whereas word
recognition was significantly correlated with reading improvement (r= 0.631; p=0.0038)
as well as phonological awareness improvement (r=0.661; p=0.0021). Finally, the mean
initial performance on repetition tasks, mainly assessing fluidity and accurateness of
speech utterance, was found correlated with the rate of improvement on phonological
awareness tasks (r=0.536; p=0.0218).
Interestingly, the Stroop interference condition was not found to be correlated with any
measure of improvement, suggesting that effects obtained with the naming and reading
conditions are independent of attentional influence.
Finally, several of E1 measurements were found inter-correlated, not shown in detail
here. One interesting result, however, the correlation between phonological awareness
and non-word repetition, is illustrated in figure 3a.
INSERT HERE FIGURE 3
To summarize, poor rapid picture naming and word recognition were predictive of
good response to phonological treatment, with an additional motor coordination
predictor for reading improvement, suggesting a possible link between the motor
automaticity component involved in the Stambak rhythm reproduction task and
participants’ responsiveness to phonological/articulatory treatment. Finally, of the 6
participants showing evidence of relative weakness in the area of motor coordination
with associated dysgraphia, none appeared to react somehow distinctly to treatment,
since they did not differ from the rest of group in terms of rate of improvement (but the
sample, here, is too small to draw firm conclusions).
As expected, all children in this study showed significant improvement in most domains
of their reading-related cognitive abilities : reading, spelling, phonological awareness,
but also a less often explored domain, that of speech production, here measured through
several repetition tasks designed to assess speech accuracy and agility. In keeping with
our expectations, derived from previous work with similar daily phonological exercises,
this general improvement may be ascribed to the effect of the auditory component of
this intervention, confirming that, even without using temporally modified stimuli, as
was the case in previous reports (Habib et al., 1999; 2002), 6 weeks of training is
sufficient to yield consistent gain in phonology, reading, spelling and speech
production. Moreover, and contrary to most similar work in the literature, daily
phonology exercises exclusively used auditory stimulation, a particularity which will be
Our decision of choosing to use unmodified speech in training exercises follows our
concern of minimising the number of different potentially active components present
within the same remediation. This choice does not mean however that effects would not
have been stronger if temporally-modified speech had been used instead, but our study
does not allow to discuss further this otherwise still debated issue (Gillam et al., 2001a,
b; Hook et al., 2001; Troia & Whitney, 2003).
Likewise, although providing feedback is usually considered of valuable support in
remediation practice, it may involve additional mechanisms (see Tricomi et al., 2006)
which one would prefer to avoid for experimental purpose. Arguably, if motivation is
an important factor for training efficacy in the “ecologic” context of remediation, it
could be important to minimize its role in experimental perspective, when trying to
obtain scientific evidence of efficacy of a specific intervention.
The main result of this study was that combining articulatory and auditory phonological
training adds significant advantage to auditory training alone, at least as concerns those
measures depending on phonological processes, i.e. phonological awareness and non-
word reading (phonological decoding). This combined effect is remarkable because,
although modest, it allows to draw valuable indications for dyslexia therapy, as well as
interesting speculations as to some theoretical issues.
Training phonology with purely auditory exercises.
As pointed out earlier, most of the previously reported successful interventions for
phonological deficit in dyslexics have dealt with more or less complex “phonics”
methods, indistinctly providing auditory, visual, verbal and non verbal stimulation, so
that it remained unclear whether positive changes may occur in dyslexics from
phonology training only using the auditory modality. Moreover, contrary to most
similar work taking advantage of the motivational content of video-games, with
systematic reinforcement and valorisation of performance, we chose to minimize this
aspect to get rid of any additional factor of uncontrolled variability.
Although there is strong evidence that dyslexics as a group suffer from some kind of
auditory perceptual deficit (see for example Banai & Ahissar, 2006 for a recent account),
there is no consensus about the causal relationship between this deficit and impairment
in reading-related processes in dyslexics (Rosen, 2003; Ramus et al., 2003). Except for
rare studies having tried to train normal (Moore et al., 2005) or dyslexic children
(Schaffler et al., 2004) with auditory stimulation, evidence is lacking of a specific effect of
phonological awareness training through exclusively auditory exercises. In a recent
meta-analysis of the literature relevant to interventions on dyslexics, Alexander and
Slinger-Constant (2004) point out the fact that the two most famous treatments targeting
auditory processing, FastForword®(FFW) and Earobics® (Pokorni et al., 2004), are not
really specific to the auditory modality: although presented through the auditory
channel, FFW also trains syntactic and semantic comprehension, and Earobics
incorporates graphemes and written words into the program.
The phonological exercises included in the present study typically draw on those
cognitive processes referred to under the term of “phonological awareness” (Goswami
& Bryant, 1990), covering a wide range of processes from phoneme discrimination to
explicit metaphonology, including implicit phoneme sensitivity, auditory attention, and
some forms of verbal working memory. Our results thus lend support to the widely
admitted notion that training phonological awareness is a helpful instrument in order
for dyslexics to successfully learn to read, even though, as often pointed out, such
training efficacy is strongly enhanced when associated with explicit teaching of
grapheme-phoneme correspondence (McCandliss et al., 2003; see Castles and Coltheart,
2004 for a discussion of this point).
Of particular interest, in this context, is the finding in the present study of a correlation
between rapid naming on pre-tests and overall gain in phonological awareness, with
slower naming performance being associated to larger overall pre-post phonological
improvement. Specifically, poor pre-training rapid naming is a good predictor for
phonological improvement, confirming that the two processes are closely related
(Vellutino et al., 2004).
Finally, our results with the rhythm reproduction task (Stambak, 1951) are remarkable,
not only because they confirm that children with dyslexia are often impaired on such
tasks (here scoring on average below 25th percentile), but also because they demonstrate,
through the outcome of correlation analyses, that this purely auditory-motor task is
strongly correlated with improvement on reading-related measures, arguing for non-
linguistic determinants to reading impairment, such as postulated in the cerebellar
theory (see below).
Usefulness of training articulation
The main result of this study was thus the demonstration of statistically significant
benefit of adjoining articulatory training to more classical phonological exercises. This
was mostly obvious on tasks relying on phonological processing, i.e. phonological
awareness and phonological decoding. For these two tasks, the advantage of combining
auditory and articulatory training was clearly demonstrated through the A-B/B-A
design, which revealed a reversed effect of group. Since both tasks require access to
phonological — especially phoneme — representation, this strongly suggests that we
have thus been acting at the level of phoneme representation. In other words,
simultaneous training of auditory and articulatory phonology appears an efficient way
to improve and/or reinforce poor representation of phonemes, a deficit considered both
common to most dyslexic readers and central to their reading difficulties (Swan &
Goswami, 1977). Therefore, it should be important for clinicians and other specialists
involved in remediation to possess such a potentially useful additional tool among their
range of possible interventions.
More specifically, our correlation study showed that training was more efficient when
initial impairment was more severe on a non-word repetition task, especially for
progression in phonological processes. We also looked at this correlation considering
separately the first and second training periods for each group, and found that this
correlation was even stronger for periods of combined articulatory and auditory
training. In addition, the correlation between non-word repetition and phonological
awareness at E1 confirms that the two processes share common mechanisms.
Taken together, these results are indicative of a close link between auditory and
articulatory mechanisms in phonology, even though articulatory treatment did not seem
to add much to background auditory training on repetition tasks.
From a more theoretical point of view, it seems appropriate to recall here that the idea of
a link between articulation and phonology during early stages of development is not a
new one. In their “motor theory of speech perception”, Liberman and Mattingly (1985;
1989) postulate the existence of a specific module within which “speech perception
processes the acoustic signal so as to recover the coarticulated gestures that produced it.
These gestures are the primitives that the mechanisms of speech production translate
into actual articulator movements”.
It is conceivable that some children with dyslexia have suffered from early disruption of
that system and that simultaneously training its auditory and motor components can
help restore harmonious interplay between parts of the system. Accordingly, the
classical McGurk effect has been found defective in children with learning disabilities
compared to normal learning controls (Hayes et al., 2003a), suggesting a specific
inability for dyslexics to properly achieve sensory motor integration between an
auditory message containing phonemic information and a visual stimulus conveying
bucco-facial features of the same phoneme.
A new look at the cerebellum theory of dyslexia
Fawcett and Nicolson (1999) have pointed out that many children with dyslexia also
experience a series of coordination and balance problems, which could suggest
cerebellar involvement. Moreover, dyslexics are also generally slow to acquire
procedures, which these authors have referred to as a deficit in automatisation. In their
formulation of the theory, Nicolson et al. (2001) suggest that a general motor
coordination deficit would impact on the development of phonology through the
suspected links between articulation and phonological awareness, a possibility
consistent with later evidence (Lalain et al., 2003) of systematic subtle deviances in
phoneme production in dyslexics. Finally, non-motor functions of the cerebellum are
more and more often emphasized in the cognitive literature, describing the cerebellum
together as a motor organ, a cognitive organ, and an organ of learning (Ito, 2005).
The correlation found in our study between motor tasks (such as rhythm reproduction
and non-word repetition) and cognitive performances and their improvement, would
fairly well take place in such a framework. Although the rhythm reproduction task we
have used also likely involves auditory and short-term memory processes, it basically
represents a measure of the quality of output realisation and this motor component
makes it a plausible candidate as an indicator of cerebellar dysfunction (Chen et al.,
In our study, the 6 patients with (moderate) coordination impairment (mainly
clumsiness, delay in motor development, associated dysgraphia), and the 2 with
suspicion of associated hyperactivity did not seem to behave differently on pre-test as
well as pre-post training improvement than the rest of the group. Moreover, they were
not especially poor at the rhythm reproduction task, which was one of the best
predictors of improvement. Taken together, these results suggest that although possibly
related to a sensory-motor component, improvement on reading-related tasks occurred
independently of the presence or absence of comorbid coordination deficits, as
suspected by Raberger and Wimmer (2003) as well as Ramus et al (2003a). Likewise, the
absence of statistical link between the Stroop interference score and rate of progression
across training sessions, does not support any contribution of associated attentional or
executive deficit.. While recognizing that this constitutes only an indirect and
speculative case for a cerebellar origin of dyslexia, we believe it opens interesting
perspectives for future research.
Limitations of the present study
A first methodological issue, also raised in previous studies (Wise et al., 1999), is that of
comparing training periods with different amounts of input (auditory vs auditory +
articulatory periods), which could weaken the main result of this study, that of a larger
effect of the combination of the two methods. However, it must be said that this
superiority of the dual treatment appears more clearly on those tasks where it was most
expected to occur, and not in most other tasks.
Also, it is important to note that, in this study, like others whose experimental designs
involve repeated measurements at relatively short intervals (here 3 weeks), some
apparent improvement may in fact only reflect practice effects. Although this could be
the case for all tasks using meaningful materials, it is much less likely to have occurred
with non-lexical materials such as non-word reading and spelling, or purely
phonological tasks. Note that our assessment phonological tasks were totally different
from those used during training. Moreover, our results on phonology and non-word
reading are particularly robust, since the advantage of auditory/articulatory training
emerged just as much whether it occurred during the first (E1-E2) or the second (E2-E3)
session, allowing us to rule out a learning effect.
One important limitation of the present study is that we have no follow-up data, so that
the improvements observed could possibly prove to be only transient. Moreover, the
effects observed are rather small, so one can wonder about the possibility of
generalizing such results. Of course, this would be of considerable importance in order
for clinicians to consider using these data in their practice, but one must keep in mind
that with the kind of experimental design we used here, we did not expect larger effects
to occur. On the contrary, in the context of such theory-based, fine-grained
interventions, even a small effect, and even on a short period of time, is an indication
that the underlying theory may apply and that the small effect detected is apt to be
amplified if associated to other tools within a wider intervention, (as is generally the
case in clinical —and even more in pedagogical— practice).
In conclusion, we would certainly recommend using auditory and articulatory
treatments in association (or temporal succession) with other techniques favouring
grapheme-phoneme mapping, with the objective of maximizing the effect by acting at
different levels of the deficit. However, for scientific purpose, we would favour
proceeding step by step, testing each component as exclusively as possible, at the
expense of effects magnitude.
Finally, now that there is no doubt anymore about the crucial role, in dyslexia, of
training phonology, further studies should concentrate on how to take into account all
the associated symptoms and/or pathologies which usually accompany the reading
disorder. Even if the gains are of moderate, even modest, magnitude, it is of theoretical
importance to try to disentangle potential factors of improvement in order, as was
tentatively shown in this article, to progressively improve the pertinence of our
remediation protocols, and, ultimately, the efficacy of our interventions in learning
Agnew, J.A., Dorn, C., Eden, G.F. (2004). Effect of intensive training on auditory processing and reading
skills. Brain Lang, 88:21-25.
Alexander, A.W., Anderson, H.G., Heilman, P.C. (1991). Phonological awareness training and remediation
of analytic decoding deficits in a group of severe dyslexics. Annals of Dyslexia, 41: 193-206.
Alexander ,A.W., Slinger-Constant, A.-M. (2004). Current Status of Treatments for Dyslexia: Critical
Review. J Child Neurol, 19(10): 744-758.
Aylward, E., Richards, T., Berninger, V., Nagy, W., Field, K., Grimme, A., Richards, A., Thomson, J.,
Cramer, S. (2003). Instructional treatment associated with changes in brain activation in children with
dyslexia. Neurology ,61: 212-219.
Banai, K., Ahissar, M. (2006). Auditory Processing Deficits in Dyslexia: Task or Stimulus Related? Cereb
Cortex, Jan 11; [Epub ahead of print]
Boada, R., Pennington, B.F. (2006). Deficient implicit phonological representations in children with
dyslexia. J Exp Child Psychol., Aug 1; [Epub ahead of print]
Castles, A, Coltheart, M. (2004). Is there a causal link from phonological awareness to success in learning
to read? Cognition. 91(1):77-111
Chen, J.L., Penhune, V.B., Zatorre, R.J. (2005). Tapping in synchrony to auditory rhythms: effect of
temporal structure on behavior and neural activity. Ann N Y Acad Sci., 1060:400-403.
Chevrie-Muller, C., Simon, A.-N., Fournier, S., Brochet, M.-O. (1997): Batterie langage oral-langage écrit-
Mémoire-Attention L2MA. Paris : Editions du Centre de Psychologie Appliquée.
Conners, C.K. (1989) Conner’s Rating Scales: CPRS-39, CTRS-39 . Multi-Health Systems, NY : North
Demonet, J.F., Taylor, M.J., Chaix, Y. (2004). Developmental dyslexia. Lancet, 363(9419):1451-60
Ehri, L.C., Nunes, S.R., Willows, D.M., Valeska Sxhuster, B., Yaghoub-Zadeh, Z., Shanahan, T. (2001).
Phonemic awereness instruction helps children learning to read : Evidence from the National Reading
Panel’s meta-analysis. Reading Research Quaterl, 36:250-287.
Elbro, C., Petersen, D.K. (2004). Long-term effects of phoneme awareness and letter sound training: an
intervention study with children at risk for dyslexia. Journal of Educationl Psychology, 96 (4): 660-670.
Fawcett, A., Nicolson, R., Dean, P. (1996). Impaired performance of children with dyslexia on a range of
cerebellar tasks. Ann Dyslexia 46 : 259-283
Fawcett, A., Nicolson, R. (1999).Performance of dyslexic children on cerebellar and cognitive tests. J Motor
Behav., 31 : 68-78
Finch, A., Nicolson, R., Fawcett, A. (2002). Evidence for a neuroanatomical difference within the olivo-
cerebellar pathway of adults with dyslexia. Cortex, 38 : 529- 539.
Foorman, B.R., Francis, D.J., Fletcher, J.M., Schatschneider, C., & Mehta, P. (1998). The role of instruction
in learning to read: Preventing reading failure in at-risk children. Journal of Educational Psychology, 90 :
Gillam, R.B., Crofford, J.A., Gale, M.A., Hoffman, L.M. (2001a). Language change following computer-
assisted language instruction with Fast ForWord or Laureate Learning Systems software. Am J Speech
Lang Pathol., 10 : 231–247.
Gillam, R.B., Loeb, D.F., Friel-Patti, S. (2001b). Looking back: A summary of five exploratory studies of
Fast ForWord. Am J Speech-Lang Pathol., 10 : 269–273
Goswami, U., & Bryant, P. (1990). Phonological skills and learning to read. London: Erlbaum.
Griffiths, S., Frith, U. (2002). Evidence for an articulatory awareness deficit in adult dyslexics. Dyslexia, 8 :
Habib, M. (2000). The neurological basis of developmental dyslexia: an overview and working hypothesis.
Brain, 123: 2373-2399.
Habib, M., Espesser, R., Rey, V., Giraud, K., Bruas, P., Gres, C. (1999). Training dyslexics with acoustically
modified speech: evidence of improved phonological performance. Brain & Cognition, 40 : 143-146.
Habib, M., Rey, V., Daffaure, V., Camps, R., Espesser, R., Démonet, J.F. (2002). Phonological training in
dyslexics using temporally modified speech: A three-step pilot investigation. International Journal of
Language & Communication Disorders, 37 : 289–308
Hayes, E.A., Tiippana, K., Nicol, T.G., Sams, M., Kraus, N. (2003). Integration of heard and seen speech: a
factor in learning disabilities in children. Neuroscience Letters, 351 : 46–50.
Hayes, E.A., Warrier, C.M., Nicol, T.G., et al. (2003) Neural plasticity following auditory training in
children with learning problems. Clin Neurophysiol.,114:673-684.
Hénin, N., DuLac, M. (1985). Lecture et orthographe. Ecole d’orthophonie Marseille (unpublished materials).
Hook, P.E., Maracuso, P., Jones, S. (2001). Efficacy of Fast ForWord training on facilitating acquisition of
reading skills by children with reading difficulties – a longitudinal study. Ann Dyslexia, 51 : 75–96
Ito, M. (2005) Bases and implications of learning in the cerebellum — adaptive control and internal model
mechanism. Progress in Brain Research, 148 : 95-107.
Ivry, R.B., Spencer, R.M., Zelaznik, H.N., Diedrichsen, J. (2002). The cerebellum and event timing. Ann N
Y Acad Sci., 978 :302-317.
Jacot-Descombes, C., Assal, G.(1986). Epreuves continues : Lecture, dénomination en condition Stroop.
Revue Suisse de Psychologie, 45 (4) : 255-270.
Kujala, T., Karma, K., Ceponiene, R., Belitz, S., Turkkila, P. (2001). Plastic neural changes and reading
improvement caused by audiovisual training in reading-impaired children. Proceedings of the national
academy of science of the United States of America, 98 : 10509-10514.
Lalain, M., Joly-Pottuz, B., Habib, M. (2003). Dyslexia : the articulatory hypothesis revisited. Brain &
Cognition 53 : 253–256.
Lanteri, A. (1995). Restauration du langage chez l’aphasique. Bruxelles : De Boeck Université.
Lefavrais, P. (1965). Test de l’Alouette. Paris : Editions du Centre de Psychologie Appliquée.
Liberman, A.M., Mattingly, I.G. (1985). The motor theory of speech perception revised. Cognition, 21 :
Liberman, A.M., Mattingly, I.G. (1989). A specialization for speech perception. Science, 243 489-494.
Magnan, A., Ecalle, J., Veuillet, E., Collet, L. (2004). The effects of an audio-visual training program in
dyslexic children. Dyslexia, 10:131-140.
McCandliss, B.D., Beck, I., Sandak, R., Perfetti, C. (2003). Focusing attention on decoding for children with
poor reading skills : a study of the Word Building intervention. Scientific Studies of Reading, 7 : 75-105.
Merzenich, M.M., Jenkins, W.M., Johnston, P., Schreiner, C, Miller, S.L. & Tallal P. (1996). Temporal
processing deficits or language-learning impaired children ameliorated by training. Science, 271: 77-81.
Moore, D.R., Rosenberg, J.F., Coleman, J.S. (2005). Discrimination training of phonemic contrasts enhances
phonological processing in mainstream school children. Brain Lang., 94(1):72-85.
Mousty, P., Leybaert ,J., Alegria, J., Content, A., Morais, L. (1994). BELEC : Une batterie d’évaluation du
langage écrit et de ses troubles. In : Evaluer les troubles de la lecture. Gregoire J., Pierart B., eds. Bruxelles
: DeBoeck Université : 127-145.
Nicolson, R.I, Fawcett, A.J Developmental dyslexia: the role of the cerebellum. Dyslexia, 1999; 5: 155-177.
Nicolson, R.I., Fawcett, A.J., Berry, E.L., Jenkins, I.H., Dean, P., Brooks, D.J. (1999). Association of
abnormal cerebellar activation with motor learning difficulties in dyslexic adults. Lancet, 353: 1662-
Nicolson, R.I., Fawcett, A.J., Dean, P. (2001). Developmental dyslexia: the cerebellar deficit hypothesis.
Trends Neurosci., 24: 508-511.
Ortiz Gonzalez, M., Garcia Espinel, A., Guzman Rosquete, R. (2002). Remedial interventions for children
with reading disabilities: Speech perception--An effective component in phonological training? Journal
of Learning Disabilities, 35(4):334-342.
Pokorni, J., Worthington, C., Jamison, P. (2004). : Phonological awareness intervention: Comparison of
Fast ForWord, Earobics, and LiPS. J Educ Res , 97:147-15
Raberger, T., Wimmer, H. (2003). On the automaticity/cerebellar deficit hypothesis of dyslexia:
balancing and continuous rapid naming in dyslexics and ADHD children. Neuropsychologia,
41 : 1493-1497.
Rack, J.P., Snowling, M.J. & Olson, R.K. (1992) The nonword reading deficit in developmental
dyslexia: a review. Reading Research Q., 27 : 29–53.
Ramus, F. (2003) Developmental dyslexia: specific phonological deficit or general sensorimotor
dysfunction? Curr Opin Neurobiol., 13 : 212 – 218.
Ramus, F., Pidgeon, E., Frith, U. (2003a). The relationship between motor control and phonology in
dyslexic children. J Child Psychol Psychia., 44 : 712-722.
Ramus, F., Rosen, S., Dakin, S.C., Day, B.L., Castellote, J.M., White, S., Frith, U. (2003b). Theories of
developmental dyslexia: insights from a multiple case study of dyslexic adults. Brain, 126:841-865.
Rey, A. (1959). Test de copie et de reproduction de mémoire de figures géométriques complexes. Paris : Les
Éditions du Centre de Psychologie Appliquée.
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, 31 : 509–527.
Schaffler, T., Sonntag, J., Hartnegg, K., Fischer, B. (2004). The effect of practice on low-level auditory
discrimination, phonological skills, and spelling in dyslexia. Dyslexia, 10(2):119-30.
Shaywitz, B.A., Shaywitz, S.E., Blachman, B.A., Pugh, K.R., Fulbright, R.K., Skudlarski, P., Mencl, W.E.,
Constable, R.T., Holahan, J.M., Marchione, K.E., Fletcher, J.M., Lyon, G.R., Gore, J.C. (2004)
Development of left occipito-temporal systems for skilled reading in children after a phonology based
intervention. Biol Psychiatry, 55:926-933.
Signoret, J.L. (1991). Batterie d’efficience mnésique. BEM 144. Collection Esprit et Cerveau. Fondation IPSEN.
Paris : Editions Elsevier.
Simos, P.G., Fletcher, J.M., Bergman, E., Breier, J.I., Foorman, B.R., Castillo, E.M., Fitzgerald, M., &
Papanicolau, A.C. (2002). Dyslexia-specific brain activation profile becomes normal following
successful remedial training. Neurology, 58 : 1203–1213.
Snowling, M. (2001). From language to reading and dyslexia. Dyslexia , 7(1): 37-46.
Stambak, M. (1951). Le problème du rythme dans le développement de l’enfant et dans les dyslexies
d’évolution. Revue Enfance, 5: 480-502.
Stanovich, K.E. (1988) The right and wrong places to look for the cognitive locus of reading disability.
Ann. Dyslexia, 38 : 154–177.
Stoodley, C., Fawcett, A., Nicolson, R., Stein J. (2005). Impaired balancing ability in dyslexic children. Exp
Brain Res., 26 : 1-11
Swan D, Goswami U. (1997). Phonological awareness deficits in developmental dyslexia and the
phonological representations hypothesis. Journal of Experimental Child Psychology 66(1):18-41.
Tallal, P. (1980). Auditory temporal perception, phonics and reading disabilities in children. Brain and
language, 1980; 9 (2): 182-197.
Tallal, P., Miller, S.L., Bedi, G., Byma, G., Wang, X., Nagarajan, S.S., Schreiner, C.J., Jenkins, W.M. &
Merzenich, M.M. (1996). Language comprehension in language-learning impaired children improved
with acoustically modified speech. Science, 271:81-83.
Temple, E. Deutsch, G.K., Poldrack, R.A., Miller, S.L., Tallal, P., Mezernich, M.M., Gabrielli, J.D.E. (2003).
Neural deficits in children with dyslexia ameliorated by behavioral remediation : evidence from
functional MRI. Proceedings of the national Academy of science Of the United States of America, 100 : 2860-
Torgesen, J.K., Alexander, A.W., Wagner, R.K., Rashotte, C.A., Voeller, K.K.S., Conway, T. (2001).
Intensive remedial instruction for children with severe reading disabilities: Immediate and long-term
outcomes from two instructional approaches. Journal of Learning Disabilities,34: 33–58.
Tricomi, E., Delgado, M.R., McCandliss, B.D., McClelland, J.L., Fiez, J.A. (2006). Performance feedback
drives caudate activation in a phonological learning task. Journal of Cognitive Neuroscience, 18
Troia, A., Whitney, S. (2003). A close look at the efficacy of Fast ForWord Language for children with
academic weaknesses. Contemp Educ Psycho., 28 : 465–494
Vellutino, F.R., Fletcher, J.M., Snowling, M.J., Scanlon, D.M. (2004). Specific reading disability (Dyslexia):
what we have learned in the past four decades? Journal of Child Psychology and Psychiatry, 45 : 2-40.
Visser, J. (2003). Developmental coordination disorder : a review of research on subtypes and
comorbidities. Human Movement Science, 22 : 479-493.
Wise, B.W., Ring, J., Olson, R.K. (1999). Training phonological awareness with and without attention to
articulation. Journal of Experimental Child Psychology, 72 : 271–304
Population characteristics : sample size, mean (standard deviation), and significance of group
Group 1 (n=10)Group 2 (n=9) difference
Reading age (mth)
WISC pict completion
WISC digit memory
PM 47 percentile
Semantic fluency 17.8 (4)p=0.898
Table 2 : Main neuropsychological characteristics : mean (standard deviation), and significance
of group differences.
Group 1 (n=10) Group 2 (n=9)difference
Raw score 60.4 (8.7)
z-score -2.2 (0.6)
Animals naming (time)
Raw score 49.8 (17)
z-score 0.1 (1.3)
1.3 (1) p=0.4034
Anim. Reading (time)
Raw score 47.8 (23.1)
z-score 2.6 (2.6)
Anim. Stroop (time)
Raw score 85.1 (19)
z-score 1.8 (1.2)
Raw scores 13.8 (3.5)
z-score -1.075 (1.5)
-1.135 (1.5) p=0.933
Table 3 : Performance on phonological awareness pre-training evaluation
Group 1 (n=10)
Group 2 (n=9)
Repetition CCV/20 6.5 (2.4)p=0.3755
Non-word repet./2010.5 (2.4)P=0.8677
Syllable inversion/109.1 (1.2)p=0.390
Phoneme inversion/10 7.7 (2)
7.3 (3.1) p=0.656
Table 4 : Main measures obtained before training (E1), after 3 weeks of training (E2) and after 6
weeks training (E3). Group 1 (G1) received phonological and articulatory treatment first (E1 –
E2) and then only phonological treatment (E2-E3) . Group 2 (G2) received the same in the
reversed order. For each series of measures, post-hoc comparisons are reported for each group and
each interval.** :p<0.05;
Mean repetition score
p = 0,00027 ** p = 0,175
G2p = 0,00152 **p = 0,251
p= 0,0258 **p = 0,165
G2 p = 0,0051**p = 0,248
Mean word and text
p = 0,0001 ** p = 0,101
G2p = 0,132p = 0,076
p = 0,0097 **p = 0,471
G2p = 0,101p = 0,0038 **
Mean reading words
p = 0,126p = 0,070
G2 p = 0,378p = 0,020 **
p = 0,035 ** p = 0,671
G2p = 0, 271p = 0, 0021 **
Figure 1 :
Diagram of the sequence of training for each group : each participant received two 3-
week periods of training, separated by a break of 2 weeks. Both groups receive daily
auditory phonological training during periods A and B, with additional articulatory
training during period A for group 1 , and during period B for group 2. Outcome
measurements are carried out at E1, E2, E3.
Figure 2 :
Mean performance of 19 children with dyslexia at 3 successive evaluations : E1=before
training, E2=after 3 weeks of training, E2 = after 6 weeks of training. Both groups
received auditory daily training all along the 6 weeks. Group 1 received additional
articulatory training between E1 and E2, group 2 between E2 and E3.
a : progression on a composite score of phonological awareness (/14)
b : progression on phonological decoding tasks (non-word reading)
c : progression on word reading tasks
**: p<0.05. Note that, for each of these three diagrams, scales are not derived from
standard scores but averaging scores to several independant tasks.
Figure 3 :
Correlation plots computed between : pre-test phonological awareness and non-word
repetition scores (a); rhythm reproduction task and pre/post-training progression of a
reading score (b); rapid naming and pre/post-training progression of phonological
awareness score (c).