Dyslexia Dysorthography Dyscalculia

The following presented a communication

Catherine BILLARD et al., Centre de référence sur les troubles des apprentissages, neuropédiatrie, Hôpital Bicêtre, Le Kremlin Bicêtre

Michel HABIB, Service de neurologie pédiatrique, Hôpital des Enfants de la Timone, Paris

Marie-Thérèse LE NORMAND et al., Physiologie et neurologie du développement, Inserm E 9935, Hôpital Robert Debré, Paris

Monique TOUZIN, Unité de rééducation neuropédiatrique, CHU Bicêtre

The following compiled a memorandum

Michel DELEAU, Psychologie du développement, Université de Rennes

Nicolas GEORGIEFF, Institut des Sciences Cognitives, Bron

Philippe MEIRIEU, Sciences de l’éducation, Université Lumière-Lyon 2

Scientific, editorial and logistic coordination

Fabienne BONNIN, Expert review associate, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Catherine CHENU, Scientific associate, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Jeanne ÉTIEMBLE, Director, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Cécile GOMIS, Secretary, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Anne-Laure PELLIER, Scientific associate, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Chantal RONDET-GRELLIER, Documentalist, Centre d’expertise collective de l’Inserm, Faculté de médecine Xavier-Bichat, Paris

Acknowledgments

Inserm and the expert group would like to express their gratitude to Stanislas DEHAENE (Neuroimagerie cognitive, Inserm U 562, Service Hospitalier Frédéric Joliot, CEA, Orsay) and José MORAÏS (Faculté des sciences psychologiques et de l’Éducation, Free University of Brussels) for rereading the work and for their advice.

Spoken language acquisition: chronological benchmarks

The neonate’s ability to learn his native language never fails to astonish. In a few years, the infant will master the complexity of the various constituents of the language. The linguistic and cognitive studies have tackled the question of speech and language acquisition by referring to a three-component system: form, content and usage. The formal aspects of the language are related to phonology (the limited set of sounds of a language which can be combined to form an infinite number of words) and the syntax (which organizes the order of the words). Content refers to the field of semantics (meaning of words and statements). Usage is a pragmatic field which studies the set of codes that regulate the speaker’s communicational intentions.

Language components

The studies of the emergence of acquisitions now enable a fairly clear concept of the very early specialization of perceived speech processing and the fast development trajectories with respect to lexicon and morphosyntax processing in the child. As of the first few months of life, the ability to perceive the sounds of speech enables the infant to discriminate, categorize the elemental sounds and recognize certain words of the language by prosody (the ‘musical’ envelope of speech with its aspect of rhythm, tempo, melody, accent and intonation). Towards the age of 7–8 months, the infant is able to recognize and memorize syllable forms of the ‘word’ type with clearly defined consonant-vowel sequences that are particular to the language. Toward age 9–10 months, the first words are acquired before the lexical explosion at about age 18 months. Assemblies of words occur at about age 24 months and grammatical expansion from about 30 months. While the between-individual variability is very great, the period from 0 to 3 years is decisive in the rapid implementation of the process of speech and language acquisition in the child.

From about age 3 years, the child learns to construct a story. This consists first in a list of various states which includes adverbs such as ‘here’, ‘there’ and ‘now’. At age 5 years, the child begins to make clearer and more varied links between the events taking place in a story. The child uses ‘and then’, ‘when’, ‘after’. The accentuation of particular events only emerges gradually. At age 10–11 years, the child still does not have narrative skills similar to those of the adult.

The set of data on the chronology of speech and language acquisition and story construction should enable professionals (teachers and clinicians) to detect asynchronous development. Early evaluation of the ability to segment, the emergence of prattling, the emergence of the first words and the text based on the stories in their conceptual and linguistic dimensions is justified on the basis of the predictive value that those abilities have with respect to learning to read and write.

Learning to read

The objective of reading is comprehension. To achieve that objective, the child must acquire a high degree of automatism in the identification of written words. By developing that skill, the child will achieve a level of reading comprehension equivalent to that of his/her spoken language comprehension.

In alphabetical writing, as in syllabic writing, two procedures enable identification of the written words: the sub-lexical procedure (phonological procedure or decoding procedure) and the lexical procedure (or orthographic procedure). The sub-lexical procedure is based on meaningless units, written syllables which, in syllabic writing, encode the spoken syllables or graphemes (consisting of one or several letters) and which, in alphabetical writing, encode the phonemes. The lexical procedure is based on units which have meaning: words. Decoding rapidly gradually becomes automatic. The procedure thus does not only refer to the slow and laborious reading of the beginner: the expert reader can very rapidly identify words that he/she does not know. Moreover, the lexical procedure is not based on a sight word strategy: the expert reader accesses the orthographic, phonological and semantic codes of written words in a few hundreds of milliseconds.

Operations involved in written word identification in syllabic writing and alphabetical writing

The reader who learns to read alphabetical writing is faced with three types of problem. The first problem is that, before learning to read, the subject has no orthographic representations except for a few words learned by heart. Similarly, before learning to read, children reader does not necessarily have explicit access to certain phonological units, in particular the phoneme, which is the minimal distinctive feature enabling differentiation - in a given language - of two words, for instance ‘cat’ and ‘mat’ in English. Taking into account the phoneme, which is indispensable in order to understand the principle of alphabetical writing and hence read that type of writing, necessitates a focus on abstract spoken language components, phonemes. In additions, those components that are difficult to access for reasons of co-articulation (the word ‘car’ is pronounced /kar/ and not /k+a+r/, this renders identification of the three phonemes composing the word difficult). The second problem with which the beginning reader is faced is related to the degree of transparency of the relationships between the basic units of writing (graphemes) and spoken language (phonemes) which vary depending on the language in question. Thus, in Spanish, unlike English, the relationships are very regular. French is closer to Spanish than to English for the regularity between the graphemes and phonemes (used to read) while French is closer to English with regard to the relationships between phonemes and graphemes (used to write). The problems of transparency with regard to the relationships between the basic units of writing and speaking are explained by the history of languages (while the spoken language evolves, the written language is more conservative) but also by the fact that the letters of the alphabet do not enable transcription of the phonemes of the various languages. For example, there are only 5 or 6 letters for vowels. This does not enable transcription of the 16 vowels in French. A combination of letters therefore has to be used or a letter to which a specific mark has been added is used to transcribe certain vowels (for example, the ‘é’ in contrast ‘è’ and ‘e’ or ‘ou’ compared ‘u’ and the letters ‘a, o, u, i’ followed by ‘n’ which indicates a nasal vowel: ‘an’, ‘on’, ‘un’, ‘in’). The basic unit of alphabetical writing is therefore not the letter but the grapheme which refers to the phoneme, the basic unit of the phonological system. The last problem is related to the size of the units. There are many more orthographic units to be learned in writing which uses large-scale units (e.g. the word) than in those which use small-scale units (e.g. the grapheme reflecting the phoneme). However, the large-scale units have a meaning and the small-scale units do not. The result is that small-scale units are less readily accessible than large-scale units³.

The research on learning to read has shown that the ease of learning depends on the degree of transparency of the grapheme-phoneme correspondences, which is very high in Spanish and very low in other languages such as English. French occupies an intermediate position. Accordingly, the worst reading scores are observed among English-speakers and the best among Spanish-speakers. For example, in France, in primary school (grade: CP, 6–7 years), the scores for reading regular words (e.g. ‘table’) and pseudo-words (regular words which do not exist, e.g. ‘tople’) do not differ and are higher than the scores observed for irregular words (e.g. ‘sept’), which are very low. In a few months (at the end of CP), the picture changes markedly. The child improves for all items but the improvement is more marked for regular words, which are better read than pseudo-words, which are in turn better read than irregular words. The foregoing is probably due to the fact that regular words benefit from both their regularity and the frequency of exposure. Neither frequency of exposure alone nor regularity alone is sufficient as shown by the weaker progression for irregular words on the one hand and pseudo-words on the other. These data enable it to be understood why Spanish children who practically only encounter regular words make rapid progress while learning to read is slow and laborious for English children.

Linguistic factors induce not only quantitative but also qualitative differences in the procedures for written word identification. For instance, the lexical procedure is more used when the orthography is opaque. This is not the case when the orthography is more transparent.

Linguistic factors induce differences in the reading units used. For example, English promotes greater use of units such as word rhymes because, in that language, the pronunciation of vowels largely depends on the consonants that follow. This is not the case in languages which have clearly pronounced vowels such as Spanish, German and French.

In contrast, the (rare) studies in the field of morphology⁴ show that, irrespective of the opaqueness of the orthography, children use morphological units when they read. However, that ability, which only seems to be acquired progressively, is dependent on phonological factors until a late period.

Lastly, irrespective of the deepness of the orthography, the grapheme-phoneme decoding ability determines the degree of success in learning to read. In fact, readers who are, from the outset, the best decoders are also those who progress the fastest, including with regard to reading irregular words and understanding text.

These findings enable other research results to be understood, particularly:

• why systematic and early teaching (as of the start of primary school, grade CP, 6–7 years) of grapheme-phoneme correspondences is the approach that most effectively helps pupils. When the method is introduced later, its impact is weaker. Moreover, that type of teaching is particularly beneficial for children who may experience difficulties in learning to read (future below average readers and dyslexics), irrespective of whether the risk factor is related to socioeconomic conditions, educational or linguistic level;
• why it is essentially training in phoneme analysis that has an effect on learning to read. The effect is, however, more marked in children at risk for reading acquisition (socioeconomic or linguistic factors) than in readers in difficulty, reflecting the fact that prevention is easier than remediation. The effect of that type of training is nonetheless more marked if, in addition, the children can handle the letters matching the phonemes;
• why the most reliable predictors of learning to read are the phoneme analysis abilities and the level of letter recognition, together with short-term phonological memory and rapid naming. Those predictors enable rapid detection (as of the end of the last section of nursery school, 5–6 years) of future below average readers and dyslexics, with a high degree of reliability. It may be considered that phoneme analysis abilities are of crucial importance at the start of learning to read in an alphabetical writing system because they enable the child to access the principle of that type of writing. The degree of intervention of the skills involved in completing tests which evaluate both the speed and accuracy of spoken lexicon access (rapid naming of images of objects, colors, etc.) is less clear. It may, however, be supposed that, initially, when the child is mainly using decoding, the ability enables rapid and precise access to the spoken word matching the series of letters decoded. This would facilitate the making of links between the orthographic code and phonological code for written words.

Learning to write and orthography

Learning to write is a relatively long process and causes problems for all children. The latter are usually overcome towards the third year of primary school in the majority of children. During that learning setup phase for all children and beyond it for those who have difficulty making writing automatic, the difficulties experienced in writing impact on the quantity and, in a less sure manner, on the quality of written language production and on orthographic performance.

In all alphabetical writing systems, including French, the crucial stage in learning consists in understanding the fact that the sequences of letters have regular matches - sometimes complex - with sound sequences, in other words the alphabetical principle. In production, this means that the child must have achieved a certain degree of mastery of the native language so that the child can use it as an object to be observed in order to, for example:

• segment statements;
• retain the segmentation in memory while transcribing, even in a non-conventional manner;
• check afterwards, by reading, the (relative) accuracy of what the child has written.

In transparent orthographic systems, access to the alphabetical principle is practically enough to ensure transcription of all new words, whatever they may be. This does not apply in French since the relationships between phonemes and graphemes are irregular.

Since French cannot be transcribed by only referring to the associations between phonemes and graphemes, lexical information specific to the words (e.g. the spelling of the word ‘thym’), general orthographic information (e.g. no double consonants at the start or end of words) and information relating to morphology (e.g. the plural of nouns is made by adding an ‘s’) need to be acquired and recalled, depending on the circumstances. This raises the question of acquisition.

From the very start of learning to read, children are able to memorize orthographic forms and at least some of their specificities without awaiting mastery of all the phoneme-grapheme matches. Little is known about between-individual differences, the number of exposures necessary in order to learn written words or the manner in which orthography becomes more conventional.

Also very early, as of the first year of primary school, children exposed to writing extract orthographic regularities which are not simply those reflecting phonological regularities or simple memorization of instances. Children are able to draw on those regularities in writing tasks, thus exhibiting very early acquisition of certain orthographic conventions which do not go so far as extracting rules, even when that would be possible. Too few of those regularities have been studied to enable determination of the conditions under which they are taken into account and the potential between-individual differences in learning them.

Traces of the use of morphology-related markers are also observed very early. However, studies in which the effects of graphotactic constraints were differentiated from those associated with morphology show that the use of morphology begins relatively late, at least if we consider that use corresponds to rules whose application does not depend on frequency. However, numerous examples show that the use of morphological markers is not initially independent of the frequency of item exposure. In addition, since morphology is correlated with other types of phonological, orthographic and semantic information, it may be that the effects attributed to morphology are in fact due to other factors. A degree of caution is therefore required in the interpretation of the data generated by those studies alone.

Rare studies on the production of derivational morphology show the persistence of the effects of graphotactic constraints despite the possibility of using an orthographic rule. Thus, even after several years of writing, French pupils do not use a rule specifying how to write, for example /o/ when /o/ is a diminutive suffix (e.g. ‘éléphanteau’, ‘lionceau’). This leads to questions on what happens when the rules are explicitly taught and on the possibility of doing so systematically, which is the case with flexional morphology.

Flexional morphology addresses a restricted number of fields: the gender and number agreement of nouns, adjectives and verbs (‘accords’), the verbal system (conjugation). It is very productive: for example, nominal agreements ending in ‘s’ apply to most nouns and adjectives. A limited number of markers for number and gender are used: ‘s’, ‘-e’, ‘-nt’. Then there are the verbal flexions which are more numerous and many of which do not have a phonological match (‘-s’, ‘-nt’) or are associated with a given phonological form (e.g. /e/ associated with ‘é’, ‘-er’, ‘-ait’, etc.). In principle, the foregoing is systematically taught. Use seems to emerge late and very dependent on the teaching. It transits through stages that may erroneously suggest that the morphological rules have been mastered early.

Teaching consists in stating rules followed by exercises to apply the rules. The pupils learn to implement a procedure which, at the beginning, calls for time and attention. Overgeneralization errors sometimes occur: for instance, adding an ‘s’ to plural verbs. Under the effects of frequent practice and teacher feed-back, the errors are eradicated and the procedure is applied more rapidly, decreasing the amount of attention required. When certain forms are frequent, they are memorized as such. The pupils thus retrieve them directly, leading, under certain conditions, to new types of error that may also occur in adults and consist, for instance, in writing ‘il les timbres’. Awareness of the agreement rules thus occurs during checking production much more than during application of the procedure.

The data on learning to write show that several processes are involved. The first process relates to learning the associations between phonemes and graphemes: in consequence, it is highly linked to learning to read even though the relationships between writing and reading require further study. The second process also involves reading and, doubtless, writing practice. The frequent exposure to written words results in two acquisitions: first, certain words are memorized and thus directly retrievable; secondly, ‘graphotactic’ regularities, i.e. associations between letters or graphemes are extracted together with the contexts in which they occur by implicit learning (without being aware of having that knowledge). Thirdly, some of the regularities may be formalized and result in statement and application of rules associated, to a variable degree, with exceptions (e.g. all the diminutives ending in /o/ are written ‘eau’, except ‘chiot’). However, such rules do not seem to be spontaneously accessible to pupils: awareness of them and generalization seems to require teaching. Fourthly, teaching the rules of flexional morphology leads to formation of procedures whose implementation requires marked attention rendering their application fragile, at least initially. Only regular and prolonged practice decreases the attention input. The practice, in reading as in writing, results in the most flexional forms being remembered and thus directly retrieved. Sometimes, this results in emergence of new errors. The availability of formal agreement rules enables production to be checked and errors to be corrected.

While the principal processes of orthographic learning have been identified, their interactions, particularly during schooling and as a function of what is or is not taught, have yet to be investigated.

Learning arithmetic

The issues relating to acquisition of and then learning elementary arithmetic are situated at three levels. Firstly, infants, like animals, are able to mobilize two different systems to process quantities and the transformations affecting them (addition, subtraction). One system is precise but only applied to small discrete sets (1, 2 and 3); the other, which can be extended to very large quantities, operates on continuous dimensions or treats as such sets of discrete elements in order to formulate an approximate evaluation in which the errors increase with the size of the quantities estimated (Weber’s law). The question of the specifically numerical character of the corresponding processes remains open as does the question of the characteristics of the representations on which the processes are implemented. The capabilities are only a point of departure but may constitute the basis of number semantics. The more complex mathematical knowledge that man has developed over his history go much beyond the foregoing and involve symbolic number systems.

Secondly, from the age of 12–18 months, infants appropriate the system specific to their culture and enabling precise determination of the number of things. The acquisition has two dimensions: first, that of the verbal naming of quantities specific to the infant’s culture (when the culture has one); secondly, the practices of counting which may be based on universal principles. Acquisition is slow and difficult for reasons that have been partially elucidated.

The verbal systems are conventional systems based on two major principles:

• the lexicon which gives a cardinal number a single and unique name (five sixteen);
• combinatorial rules enabling elaboration of an infinity of complex formulations enabling naming of any cardinal number (six hundred seventy five million three hundred ten thousand two).

The rules enable combinations of the additive (one hundred three) or multiplicative (three hundred) types. The French spoken numerical system lexicalizes cardinal numbers up to sixteen, the tens from twenty to sixty, hundred, thousand, million and billion. Combination codes exclusively additive relations up to 79 (twenty five = twenty + five), then additive and multiplicative relationships (four hundred six = four × one hundred + six).

Quantities may be evaluated overall or precisely. In the first case, as previously indicated, the errors increase with the size of the collection to be evaluated: the errors are more rare for small collections (1, 2 or 3, or even 4) for which a specific processing mechanism may exist (subitizing). Subsequently, the errors increase. In the second case, counting is necessary, i.e. identifying each of the entities in the collection and processing it once and only once and assigning to it a verbal label and only one verbal label such that the last item’s label is the cardinal number of the collection. Counting correctly requires attention, knowledge of the numerical lexicon and the ability to implement motor skills and coordinate them with memory retrieval of the verbal forms. Using counting, children manage to assign a single and precise cardinal number to each collection, irrespective of the perceptual form of the collection.

Thirdly, infants identify the changes in quantity associated with addition, subtraction or even fractionation very early. Mastery of counting enables children to precisely quantify those changes. They can therefore implement an action which is equivalent to an arithmetic operation although it has yet to become one. They do so by grouping or separating sets and by counting both the initial sets and the results of the changes. Only very gradually will they move from resolution by action to processing addressing the symbols only. It is possible that use of the fingers during that phase plays an important role insofar as the sets of fingers are both analog and abstract (they are a substitute for all objects). Progress is reflected by the switch from external actions to interior actions, then processing conducted on verbal forms (4+3 → 4, 5, 6, 7). It is not easy to determine how infants realize that using the symbols alone yields results that are as reliable as those obtained by manipulation.

Fourthly, when the child goes to school, he/she is systematically taught a new code – the Indo-Arabic code – and the algorithms associated with it. This gives the resolution of the operations a power that the verbal code alone cannot ensure. Learning is both easy (the Indo-Arabic code only has ten items: 0,…9) and difficult, in particular due to the positional notation (the value of a digit changes with its position). Trans-coding, switching from the spoken to the Indo-Arabic code or the opposite, is initially based on verbal knowledge. This explains why, for instance in French, the transcription of quantities such as seventy five (soixante-quinze) gives rise to errors such as 6015. The algorithms concern the resolution of complex additions, subtractions and multiplications and, to a lesser degree in elementary school, divisions. The children must memorize certain associations between operands and results (3 + 2 = 5) so that they no longer need to calculate them and can devote their attention to managing the algorithm. The algorithms also require a minimum of attention to and mastery of space. The algorithms need systematic and vigilant teaching in order to prevent the installation of procedural errors that are difficult to eradicate. The introduction of fractions and decimals raises new problems since, first, the representation of quantities changes (although fractionation of quantities occurs early) and, secondly, the operation processing algorithms differ from those applied to natural whole numbers (addition of fractions, addition of decimals, etc.).

Fifthly, the resolution of arithmetic problems remains a major problem as is shown by national and international evaluations. One of the difficulties is associated with reading comprehension more than the arithmetic processes themselves, at least at elementary school. Another difficulty is linked to the ‘arithmetization’ of situations. The children have to elaborate from a statement a representation of the situation described, then transition from that representation to an arithmetic formulation. Depending on the formulation, the familiarity with the field, the number of units involved and the frequency of exposure to the situations, the elaboration of that representation and its ‘arithmetization’ may be very difficult and lead to an erroneous interpretation and hence incorrect resolution. The essential question is thus that of finding situations enabling pupils to frequently encounter a variety of problem situations so that they familiarize themselves with the diversity of statement presentations, the diversity of the situations described and, lastly, the diversity of the modes of resolution.

In short, doubtless from birth, children have the quantitative processing capacities that are biologically determined. Subsequently, cultural practices are added. The latter exploit and amplify the initial capacities. School systems install new learning systems: knowledge (decimals), activities (problem solving), codes (Indo-Arabic code, operators) and processing procedures (problem solving algorithms). Each of the dimensions is associated with particular problems, some of which remain imperfectly understood.

Definitions and classifications of specific learning disabilities

The disorders addressed herein are those related to learning in school: disorders of reading, writing and arithmetic. The disorders may emerge in a context of overall retardation or more specific retardation in the event of deficiencies restricted to certain cognitive processes. The following will mainly address the definitions of and criteria for ‘Specific scholastic learning disabilities’ present outside of overall developmental retardation.

The disorders are defined in the classifications as a set of learning difficulties that cannot be attributed to intellectual retardation, a sensory handicap or unfavorable environmental conditions. The difficulties are thus unexpected given the other aspects of development. The difficulties emerge very early in life, interfere with integration in school and social integration and frequently persist until adulthood.

The main criteria for specific learning disabilities (in reading, writing or arithmetic) which constitute the bases for the definitions proposed in the first years of life are as follows:

• the criterion of a ‘discrepancy’ between the difficulties experienced in tests related to the disorder in question and the good performance in other cognitive tests (frequently the IQ⁵);
• exclusion criterion: the primary cause of the disorder is not to be overall retardation, sensory handicap, educational deficiency (inappropriate teaching, inadequate socioeconomic level), a linguistic difficulty or mental disorder that has been clearly demonstrated;
• the disorder is due to factors intrinsic to the child (this point is directly derived from preceding two points and stresses the neurobiological origin of the disorders).

The criteria are used in the International classification of diseases, ICD-10⁶ and in the fourth edition of the Diagnostic and Statistical Manual of mental disorders (DSM-IV)⁷. Both classifications are very widely used at international level.

In line with a recent trend, the label ‘specific learning disabilities’ should not be allocated simply on the results of the various tests but should be restricted to children whose disorders are recalcitrant to appropriate treatment (pedagogic and/or speech therapy management). Moreover, progress in knowledge in recent years has led to the inclusion in the definitions of the neurobiological or even hereditary etiology of specific learning disabilities and their cognitive mechanisms even though those cognitive processes remain the subject of discussions and investigation. It should also be noted in the context of the research work that the criteria adopted may vary as a function of the number of subjects observed. Studies on large populations use, for feasibility reasons, criteria that are less rigorous and numerous than those used in studies on small groups of children.

Reading disorder or dyslexia is by far the most widely explored specific developmental disorder of scholastic skills. Generally speaking, it presents as a difficulty in learning to read despite appropriate teaching, adequate intelligence and a good socioeconomic environment.

In the event of concomitant disease, the diagnosis of specific reading disorder predominates over the other specific disorders (arithmetic and writing disorders) in ICD-10 while DSM-IV enables several diagnoses to be formulated.

Recent definitions of dyslexia address disturbed cognitive mechanisms (often a deficit in the phonological component of language) and the etiology of the disorders (neurobiological etiology).

The specific disorder of arithmetical skills or mathematics disorder or dyscalculia usually refers to a disorder of mathematical skills in children of normal intelligence. Despite indubitable progress in studies in recent years, the disturbances of cognitive mechanisms underlying dyscalculia continue to be studied even though numerous subtype classifications of dyscalculia have been proposed. The etiology is already at the core of the initial definitions of dyscalculia proposed, as are the genetic or congenital etiologies of the disorders.

In the definition of dyscalculia, as for the other specific learning disabilities, inappropriate teaching constitutes an exclusion criterion.

Since writing is a multidimensional process, several components may be responsible for the writing disorder: motor-related writing difficulty; written task completion difficulty; inadequate spelling; problems of writing composition such as planning, choice of words, sentence construction and text organization. The term dysgraphia, which rather suggests a handwriting disorder equally affecting the shaping of geometrical shapes, has sometimes been used to represent a general disorder in writing. However, in the literature, reference is more frequently made to dysorthography.

Dysorthography is a specific disorder of spelling which accompanies dyslexia; the cognitive dysfunction underlying the two disorders is probably common to both. In dysorthography, the spelling of words is highly deficient, a direct consequence of the phonological disorder in dyslexic children.

There is thus no single definition of writing disorder and this is clearly illustrated by the international classifications: the ICD-10 definition is that of dysorthography while the DSM-IV definition rather covers a more general disorder in written expression.

While insufficient and open to criticism, the international definitions and classifications (ICD-10 and DSM-IV) have the advantage of enabling comparison of studies conducted internationally and use of a common language. However, in the field, the international classifications are frequently ignored by both clinicians and researchers, who use their own classifications.

Thus, the choice of the cutoff defining the disorder (from 1 to 2 standard deviations below the mean) depends on the prevalence of the disorder. If 1 standard deviation is selected, in a normal distribution there will be 16% of subjects below the cutoff. If 2 standard deviations are selected (as proposed by the ICD-10 international classification), there will no longer be more than 2.5%. It may therefore be considered that the disorder is an entity that is in part arbitrary since dependent on the cutoff selected a priori. Nonetheless, in clinical practice, follow-up of each child and the use of numerous tests enable a real diagnosis to be formulated.

The questions raised by the concept of ‘discrepancy’ and the fairly frequent coexistence of several learning disabilities have prompted certain authors to question the concept of specificity. The DSM-IV classification, issued later than ICD-10, does not always use the term ‘specific’ without, however, changing the definitions of the disorders. It is considered important to compile a classification of learning disabilities based on common criteria shared by researchers, clinicians and educational professionals.

Prevalence data on specific scholastic learning disabilities

Operational diagnosis of ‘a case’ requires several investigations and numerous tests repeated over several months or even several years. That approach cannot be used in the context of epidemiological studies of the overall population. Moreover, in epidemiological studies, the criteria may vary between authors, as may the cutoffs selected. It is therefore not surprising that the published prevalence data vary widely. Nonetheless, on the basis of the most methodologically rigorous international studies, mainly from English-speaking countries, it may be considered that dyslexia affects between 3 and 5% of children aged about 10 years.

Studies using the same methodology to compare populations with different languages have enabled evidencing characteristics specific to each language. Languages have different writing systems and, at present, variations in the form of dyslexia and its prevalence are considered dependent on factors such as transparency (regularity) or opacity with respect to the spelling (orthography) of each language. Thus, a study using three different definitions of dyslexia showed that the prevalence varied, as expected, as a function of the definition and strictness of the discrepancy criteria but also as a function of the language. Depending on the definition, the prevalence ranged from 3.6 to 8.5% in Italy and from 4.5 to 12% in the United States. The higher prevalences in the United States, compared to Italy, are related to the orthography: Italian spelling is transparent, unlike English spelling.

In France, in the absence of epidemiological data, the data available and complying with the criteria for characterization of dyslexia suggest that the proportion of children presenting with patent dyslexia is not different from that reported in the large-scale studies of English-speakers.

Although not all reading disorders, even when they are severe, may be considered dyslexia, the data generated by two recent epidemiological studies conducted in adults in France yielded results compatible with the estimates above: one study was conducted on young military recruits and the other on 10,000 adults aged 18 to 65 years (Insee). The studies found that 7% of adults aged 18 to 29 years experienced severe or marked difficulties in reading.

Questions may be asked with regard to the potential interactions between the school environment or educational level and the biological factors of dyslexia, and with respect to their effects on the manifestations of specific learning disabilities. By way of an example, studies conducted on the English populations of the Isle of Wight and London reflect geographic and social differences. Using the same definition of dyslexia, the prevalence was 3.6% on the Isle of Wight and 9.3% in London, where the subjects were generally of lower socioeconomic status.

The prevalence of dyscalculia is frequently considered to be equivalent to that of dyslexia. However, the most rigorous studies suggest that dyscalculia, as an isolated disorder, is rarer than dyslexia. To the authors’ knowledge, no epidemiological study of dyscalculia has been conducted in France. There are few reasons, however, for thinking that the prevalence of dyscalculia differs markedly from that in other countries; the language seems to have little impact on the disorder as the convergence of the results generated in different countries shows.

Gender is one of the variation factors frequently suggested to be associated with specific learning disabilities. In general, the male/female sex ratio is estimated to be 2 for dyslexia while it is probable that dyscalculia affects girls as frequently as boys. It has also been observed that more boys than girls consult specialists for reading or writing disorders. However, the reality of a gender effect on specific scholastic learning disabilities remains a subject of debate and the reasons for any difference between boys and girls have yet to be investigated.

Longitudinal studies show that dyslexia persists with age and thus consists in a non-transient developmental deficiency even though long-term follow-up of dyslexic children raises the question of classification stability over time. Since the classifications are based on a cutoff, small changes in scores over a continuum may be sufficient for certain children to move from one side of the cutoff to the other.

Longitudinal studies also show dyscalculia to be a persistent disorder. However, the isolated forms of dyscalculia (not accompanied by dyslexia) appear to be less persistent and are considered by some authorities to be more related to developmental retardation than to a real individual difference.

In numerous studies, specific learning disability is considered an entity of genetic and cognitive etiology. It is regrettable that there are only a few studies on the interactions between genetic, cognitive and environmental factors and on the influence of the methods used to teach reading.

Dyslexia: studies of groups and multiple cases

In order to understand a text, the child must acquire a high degree of automatism in the identification of written words. Development of that skill enables the child to achieve a reading comprehension level equivalent to that of his/her spoken language comprehension by freeing the child from slow and laborious decoding or recourse to risky contextual anticipation. One manifestation of dyslexia is in fact the inability to develop that type of skill.

The review of the data on dyslexia took into account studies conducted in various languages (including English, French and German). The studies were conducted either on undifferentiated groups of dyslexic children or on groups of dyslexic children with different profiles.

The results of the studies on undifferentiated groups of dyslexic children showed, first of all, that the phonological deficiencies were encountered in almost all the studies and in almost all the subjects both in reading and outside of reading.

For instance, the reading performance of dyslexic children is particularly deficient when the children cannot exploit their lexical knowledge, i.e., when they have to read rare words or non-existent words (pseudo-words). The deficiency in the sub-lexical (phonological) reading procedure has been detected in a convergent manner in almost all the studies, including comparisons with younger children with the same reading skill level. In addition, the deficiency in the sub-lexical reading procedure is more severe when the children are confronted with opaque writing (e.g. English compared to French). Lastly, when the orthography is transparent, the deficiency in the sub-lexical reading procedure is mainly observed in processing speed, and not in accuracy score.

In addition to the sub-lexical (phonological) reading procedure deficiency, deficiencies in the lexical reading procedure are also observed in children with dyslexia. This can be understood in the light of studies that have shown that the setup of the lexical procedure depends on the efficiency of the sub-lexical reading procedure. In consequence, children most frequently suffer from a double reading deficiency; the deficit of the sub-lexical reading procedure is, however, more severe than the deficit of the lexical reading procedure. The former is in fact the only deficiency observed compared to younger normal readers with the same reading level.

Dyslexic children also have particularly deficient skills with respect to tasks that involve phonological processing outside of reading: phoneme analysis; phonological short-term memory and naming tests which enable evaluation of the precision and rapidity of lexical accessing. As was the case with reading deficiencies, the deficiencies in those three fields have been reported in most of the studies and for most dyslexic children, including in the studies vs. younger children with the same reading skill level. Lastly, the skills in those three fields are the most reliable predictors of the future reading level of children. In comparison, the weight of non-verbal skills and sociocultural factors is less. The deficiencies observed in those three fields may interfere with the setup of the reading phonological procedure. In order to use that procedure, the written code must first be translated to the spoken code, which, in an alphabetical writing system necessitates associating graphemes with the corresponding phonemes. The units resulting from that decoding operation must then be assembled in order to access the words stored in the spoken language lexicon. The first operation requires phoneme analysis skills while the second involves the short-term phonological memory and the third precise and rapid access to the spoken language lexicon.

Other studies, also conducted on undifferentiated groups of dyslexic children, addressed the deficiencies in visual sequential processing. The results published to date are not conclusive for three main reasons. First, that type of deficiency is more marked for pseudo-words than for words or for unpronounceable sequences of letters. In addition, the deficiencies in visual sequential processing may be the consequence of dyslexic reading difficulties. Indeed, in almost all the studies that have drawn attention to that type of deficiency (unlike those having pointed to phonological deficiencies), the dyslexic children were only compared with average readers of the same chronological age (not with average readers of the same reading level). Lastly, in most of the studies, the children’s visual skills were evaluated by taking into account processing speed and/or in terms of tasks involving time constraints (very brief exposure to the stimuli). In contrast, only accuracy scores were were examined for the assessment of phonological skills, and not processing speed. The results may therefore be explained by differences in measures. One of the rare studies in which visual sequential processing deficiencies and phonological processing deficiencies were intensively evaluated using a comparable methodology showed that dyslexic children who suffer selectively from visual deficiencies of that type are few in number.

A number of recent studies have evaluated a new hypothesis according to which the typical slowness problems observed in non-English-speaking dyslexic children are explained by their difficulty in memorizing the visual form of words, while the response precision deficiency observed in English-speaking children is related to a phonological deficiency. It is, however, difficult to imagine that the dyslexic phenotype may be fundamentally different depending on the transparency of the writing system and the measurement used. Moreover, the results of those studies, based on analysis of eye movements, do not enable validation of the new hypothesis given that the dyslexic children were compared to average readers of the same chronological age: the specificity of their eye movements may therefore be a consequence of their reading difficulties. Lastly, some of the results of those studies are compatible with the phonological hypothesis: for instance, the negative impact of the opacity of the orthographic system on the duration of ocular fixation; the fact that the most marked differences between the dyslexic and average reader children were observed with respect to pseudo-words; and the presence, in certain of those studies, of phonological deficiencies in future dyslexics, compared to future normal readers, before learning to read.

In addition to the studies of undifferentiated groups of dyslexic children, other studies have examined subjects considered to present differentiated dyslexic profiles such as phonological dyslexia (characterized by a selective deficiency in the phonological reading procedure, among other characteristics), surface dyslexia (characterized by a selective deficiency in the lexical reading procedure, among other deficiencies) and mixed profiles with a double deficiency. The published studies show that practically all dyslexic subjects have (or have had) a phonological deficiency, including the subjects presenting a surface profile. That is evidenced by longitudinal studies in particular. Moreover, in the comparison with younger average readers with the same reading level, only the proportion of phonological dyslexic children remained high, suggesting that most surface dyslexics simply suffer from a developmental lag, since they behave like younger normal readers. On the other hand, the phonological dyslexic children have, in the majority of cases, an atypical developmental trajectory.

The compensatory strategies developed by dyslexics may, however, mask certain realities. In particular, the expert reader has been shown to identify written words by a quasi reflex action which is little influenced by the contextual information. The least accomplished readers, particularly dyslexic subjects, use the context more in order to identify written words. It is likely that such strategies enable dyslexic subjects to overcome their phonological deficiency. The fact that dyslexic subjects read words they have already encountered better than new words is probably the result of compensatory strategies. Indirect evidence of the progressive adoption of compensatory strategies has been provided by the data generated by longitudinal studies. Most older dyslexic subjects in whom no phonological disorder could be demonstrated in fact suffered from such disorders in the early periods of their development.

In conclusion, the studies reviewed indicate that a phonological deficiency is the major characteristic of dyslexia. The fact that that type of deficiency is the most frequently reported, including compared to younger normal readers with the same reading level, is the sign of atypical development. In addition, the opacity of the orthography is an exacerbating environmental factor. In the current state of the research, the evidence in favor of a visual deficiency underlying dyslexia (or certain forms of dyslexia) remains fragile.

Dyslexia subtypes: single-case studies

Most of the neuropsychological and neuroscience or genetic studies on the specific disorders of reading acquisition concluded that the dyslexic population was highly heterogeneous. However, the heterogeneity has not been evidenced in the context of studies of groups designed to detect the characteristic general trends of the dyslexic population without taking into account the variability observed in that population. Numerous case studies have nonetheless been published and reflect the diversity of the forms of dyslexia. The studies not only enable the cognitive profile of a given individual to be defined but also the cognitive component(s) whose dysfunction accounts for the deficient performance observed. The studies are therefore particularly pertinent for healthcare professionals confronted with specific cases and needing to formulate hypotheses on the nature of the underlying cognitive deficiency on the basis of neuropsychological assessment data with a view to orienting therapeutic strategy. In theoretical terms, case studies are indispensable for validation of the theoretical models. They may also enable identification of new cognitive components that are potentially involved in dyslexic disorders and whose pertinence can subsequently be tested in larger populations.

Several clearly distinct forms of dyslexia have been observed in children. Although very rarely described, the peripheral development dyslexic forms - visual dyslexia (essentially characterized by preferential production of visual errors) and negligence dyslexia (errors on the first or last part of the word independently of its linguistic characteristics) – suggest that disorders of visual processing at attentional component level may interfere with learning to read.

Phonological dyslexia is characterized by a selective disorder in reading pseudo-words and the production of phoneme errors. Phonological dyslexia is accompanied by dysorthography of the same type characterized by weak performance in taking dictation of pseudo-words. All the cases reported in the studies presented with an associated disorder of meta-phonological capability and several demonstrated a deficiency in short-term verbal memory. Those forms of dyslexia therefore seem to be associated with a phonological dysfunction probably underlying the children’s learning disabilities. Case studies of phonological dyslexia thus corroborate the conventional hypothesis that phonological deficiency gives rise to dyslexic disorders. However, certain prototypical cases call for modulation of the concept that phonological skills enable, via the analytical procedure, establishment of memorized orthographic representations. Cases showing excellent performance in reading regular and irregular words despite a massive phonological disorder and a highly deficient analytical procedure suggest that mechanisms of another nature also contribute to memorizing lexical knowledge.

In rare cases, a child may present with an extreme profile of surface dyslexia with no phonological deficiency. The dysorthography is massive: the words are written as they are pronounced without taking into account their orthographic characteristics. The children seem not to have any memorized information on the orthographic sequences of words. In that type of case, other dysfunctions may be suspected, for instance visual and attentional dysfunctions.

Mixed dyslexia is characterized by difficulties with regard to all the types of item. Mixed dyslexia has been very little studied despite its frequency in the dyslexic population. Two cases described suggest a dysfunction that is either phonological or visual and attentional may underlie this type of disorder. This does not exclude the possibility of encountering two associated deficiencies in certain subjects.

Disorders of learning to write

While the French orthography is particularly difficult to produce (but not to perceive) and it has social value and is socially defended, few in-depth studies on the difficulties associated with it and the disorders affecting it are available. Review of the few studies available and conducted with no attempt at standardization clearly shows the existence of ‘fragility zones’ ‘inducing’, in a manner of speaking, the errors. However, the mechanisms involved in error occurrence and the interventions liable to reduce error frequency have been very little elucidated or studied.

The literature available on the difficulties and disorders of orthographic production is rare with respect to written French. Currently, it does not enable rigorous differentiation of the errors encountered in the various dimensions involved in mastering orthographic production. In consequence, it is currently impossible to determine whether differentiated error patterns exist or whether certain error patterns characterize certain handicaps. The only sufficiently numerous data available are for dyslexic subjects. Once again, the data remain general (to the authors’ knowledge, no study has addressed error types).

The data available were thus mainly generated by studies conducted on other orthographic systems. Acquisition of the alphabetical principle is problematic but possible as shown by the data generated by transparent orthographic systems (Spanish, Italian, etc.): dyslexic children manage to write words. The difficulties are greater with opaque systems due to the confusion induced by the phoneme-grapheme associations (one phoneme may have several graphemes). Opaque systems do not only encode the phonology but also the lexical and morphological dimensions. Some subjects have difficulty compiling an orthographic lexicon. Sometimes, the lexicon is very limited. Sometimes the difficulties are related not to lexicon size but to its approximate nature or the erroneous nature of the memorized orthographic forms. To the authors’ knowledge, no data on the distribution of those scenarios are available. Similarly, no data on their prevalence or concomitant morbidities are available. The extraction of orthographic regularities (double consonants, frequent associations of letters, etc.) has not (yet) given rise to public research. It is therefore not known whether certain disorders are particularly involved. Difficulties relating to use of morphology affect all the population to various degrees: even the leading newspapers contain agreement errors made by eminent authors. The frequency of those errors in the overall population is not known; nor is their distribution which would enable disorders to be suspected in given subjects and related to other characteristics of the subjects involved. Since those errors are also very sensitive to the degree of attention, it is not known whether the errors occur because of misunderstanding, incorrect knowledge or implementation difficulties (for instance because of the attention load associated with writing in young children). Almost everything has still to be done.

Overall, in the absence of precise data on the errors committed by the overall child or adult population, it is difficult to determine the degree to which the number and nature of errors reflect normal performances requiring, for example, more frequent and regular practice of certain exercises or, on the contrary, constitute the grounds for suspecting a disorder necessitating specific management. This is no doubt the cause of two facts that emerged from the review of the literature on orthographic production difficulties and disorders. First, no study of the prevalence of writing learning disabilities is available: the study of those disorders is almost systematically associated with study of reading disorders (dyslexia). At least a priori, that situation leads to other hypotheses being neglected. Secondly, the studies addressing potential associations (concomitant diseases) with disorders other than reading disorders are extremely rare: one study with arithmetic disorders, a few studies with attention disorders. It should be added that, in contrast to what applies for other fields of learning, certain specificities of French orthography (the rareness of spoken language morphological markers and their systematic nature in the written language) prohibit waiting for the indispensable research to be conducted on other writing systems, enabling the solutions to the problems to be, in a manner of speaking, imported. In other words, if we wish to maintain the state of French orthography, it would appear urgent to make a radical change requiring political decisions and to address the nature and frequency of the difficulties, in a non-normative context, and the mechanisms involved in learning and implementing the various components of orthography. It is equally urgent to address the characteristics of the teaching given and evaluate its effects in the short and medium term. Lastly, prevalence and concomitant disorder studies should be conducted without considering, a priori, that the difficulties and disorders of orthographic production are, if not uniquely, at least preferentially observed in dyslexic children. Even if that is true, it is not possible to exclude the existence of disorders in other sub-populations.

In short, the difficulties involved in French orthographic production are such that one would have expected that research on those difficulties, their distributions, and underlying mechanisms and on the effects of the teaching given, including in sub-populations presenting with specific disorders, would be numerous and precise. The data show that that is not the case. The rare data are recent. Moreover, the data mainly cover descriptive aspects: study of cognitive mechanisms and their determinants, and, above all, studies of learning are still largely lacking.

Dyscalculia

Studies of dyscalculia are only beginning and our knowledge of the field cannot be compared to that accumulated with regard to dyslexia. There is no definition or diagnostic criterion universally accepted for dyscalculia. Most frequently, children whose performances are markedly lower than the mean for children of their age in a standardized assessment evaluation and when the intellectual level does not deviate excessively from the normal are considered to present with dyscalculia. These criteria do not enable dyscalculia to be distinguished from what other authors call learning disabilities in mathematics. The analysis of the prevalence studies suggests that while children presenting with difficulties in arithmetic are approximately as numerous as those presenting with difficulties in reading, specific arithmetic disorders are more rarely encountered than isolated reading disorders. In more than one case out of two and even in children with normal intellectual quotients, the difficulties in arithmetic are accompanied by difficulties in reading. In contrast to what seems to hold for dyslexia, dyscalculia affects boys and girls equally.

The broadest consensus addresses the description of the disorder and sectors of arithmetic activities showing the greatest deficiency. Well before the start of systematic learning, children presenting with dyscalculia develop a poor understanding of the principles governing counting activities (counting by pointing with the finger at a set of objects), which constitute the basis on which all subsequent arithmetic skills will be built. Such children are distinguished from others by a more frequent and more prolonged use over development of immature counting procedures to accomplish simple arithmetic tasks. Above all, children presenting with dyscalculia are distinguished by atypical difficulties in memorizing arithmetic facts and learning addition and multiplication tables. Such children have recourse less frequently than others and in a less sure manner to direct retrieval of the result from the memory when doing arithmetic problems. The difficulty proves surprisingly persistent over those children’s development. The disorders in the elementary activities have repercussions on the resolution of complex problems and operations. In the latter, use of carryovers remains difficult. The difficulties are all the more marked when the children also present with reading difficulties.

There is also agreement on the course of the disorder. Although few studies have addressed the issue, they all report that dyscalculia is a persistent disorder. However, the ‘pure’ forms in which arithmetic difficulties are the most isolated are also the most unstable, particularly in infants joining primary school. Thus, certain authors consider the isolated forms as developmental retardation rather than a true between-individual difference. During development, with the exception of the memory difficulties with respect to numerical facts, children presenting with dyscalculia manage to catch up with their peers with regard to the simplest arithmetical operations (additions).

In contrast, there is no consensus with regard to the cognitive profile accompanying dyscalculia. Apart from a few exceptions, most authors report low working memory capacities in children presenting with dyscalculia. Dyscalculia is frequently considered to also be accompanied by deficiencies in visual and spatial terms, although there is no unanimity on that question. However, those aspects are frequently deficient in children presenting with dyscalculia, but it has not been clearly established or universally recognized that they are specific to, or more pronounced in, that disorder. Children presenting with dyscalculia frequently have written language disorders affecting both reading and writing. Children presenting with dyscalculia also have attention disorders more frequently than others.

The uncertainties regarding the cognitive profile may be explained by the existence of several different subtypes of dyscalculia. Although several classifications have been put forward, they are sometimes markedly different from each other. Rarely founded on in-depth analysis of large populations, the pertinence of those classifications is dubious and their number argues against their validity. Even the most intuitively attractive classifications or distinctions have been compromised by rigorous analyses: for instance, it does not seem to be any qualitative differences between the difficulties encountered by children presenting with specific arithmetic disorders and those with associated reading disorders.

Similarly, the causes of dyscalculia have yet to be elucidated. It has not even been established whether dyscalculia is a primary disorder that can exist in an isolated manner or whether it is a manifestation of a more general disorder affecting a wider range of functions and activities. Among the general disorders which are reported to induce dyscalculia, low working memory capacity and a dysfunction of the right hemisphere inducing visual and spatial skill deficiency have been suggested. A more recent hypothesis suggests that dyscalculia may result from dysfunction of the cerebral structures specialized in numerical processing. Derived from evolution, those structures confer on human beings a ‘sense’ of numbers and geometric relationships, which are missing in subjects presenting with dyscalculia. This attractive hypothesis nonetheless requires empirical evidence. Although we are not aware of the causes of dyscalculia, all the hypotheses put forward are in agreement on the fact that dyscalculia emerges markedly before the child begins systematic learning in primary school.

Studies of remedial interventions and programs are only just beginning. Addressing small populations, the studies are in general less well controlled than studies investigating the manifestations or causes of the disorder. Although some successes with interventions have been reported (children presenting with dyscalculia taking part show an improvement in arithmetic performance), it is still too early to pronounce on their respective merits and values.

In short, our knowledge of dyscalculia is fragmentary and uncertain. This is due to the limited number of studies compared to dyslexia, for example, but also to the amplitude and difficulty of the subject of study. Number and arithmetic, not to mention mathematics, cover very diverse activities involving a large number of different cognitive functions. However, the difficulties associated with studying dyscalculia may be offset by the fact that it is not only a learning disorder. As cognitive psychology has shown, in human beings, numerical activities and their understanding develop spontaneously. This means that before the child’s first systematic learning experience, he/she has built a repertory of knowledge and know-how with regard to number and its uses. This characteristic makes possible not only the early detection of children who present with risks of subsequent learning disabilities, as is the case for reading, but also those who show atypical development in numerical skills, before any scholastic learning. It may be possible to remedy the state before dyscalculia emerges as a specific learning disability.

Associated acquisition disorders

A shared characteristic of dyslexia, dysorthography and dyscalculia is that the entities are frequently associated in a given subject. For the clinician, rehabilitator or teacher, this provides evidence of the severity of the disorder and hence its repercussions in terms of rehabilitation and teaching. But the existence of disorders associated with dyslexia also constitutes a powerful theoretical tool for the researcher in that it provides potential research orientations.

In the context of a learning disability reference center, it is not surprising to observe a marked predominance of dyslexia or severe disorders of written language acquisition (since, theoretically, we can only term the disorder dyslexia after a certain duration of learning).

Isolated dyscalculia rarely gives rise to consultation of a reference center. This means that the real frequency of dyscalculia in the overall population is higher than that indicated. Dyschronia (disorder of temporal orientation) is less known than the other syndromes, doubtless, once again, because the disorder does not, at least at first sight, constitute a real handicap with regard to scholastic progress. In any event, dyslexia is more frequently associated than isolated (only 10% in the population presented). However, there are no data enabling evaluation of the prevalence of concomitant disorders in the dyslexic population.

Dyslexia is, in over half of the cases, subsequent to spoken language disorders which have a variety of presentations. The term ‘dysphasia’ used in France for severe spoken language disorders is rarely used in the literature in English, which considers all spoken language disorders to be specific language impairment (SLI) with no preconception as to their severity.

Numerous children of normal intelligence experiencing difficulty in learning arithmetic also present with difficulties in learning written language. Certain authors have reported that in a population of children presenting with dyscalculia, 17 to 64% presented with dyslexia. However, dyscalculia also exists as an isolated disorder. It would appear that children presenting with specific difficulties in arithmetic are to be distinguished from those presenting with difficulties in the two fields. Most of the studies distinguishing the two populations observe differences in the severity of the disorders, the differences in the nature or extent of the difficulties being less certain. Most authors agree with the finding that children presenting with concomitant disorders together with reading disorders have a more marked handicap in arithmetic and in neurological test performance than children presenting with dyscalculia alone. However, the reasons for the frequent association of arithmetic and reading disorders remain obscure. In numerous studies, groups with associated disorders showed a lower intellectual level (at least descriptively and sometimes statistically significantly) than that of the groups with a simple deficiency.

Coordination disorders, which account for about 6% of the overall population with a marked predominance of boys, consist in various symptoms that are easy to recognize but are grouped together in a very variable manner. Broadly speaking, ‘praxis’ disorders, i.e. disorders in the ability to chose, plan, sequence and execute a movement, with consequences of variable degree on everyday life, are recognized. The incidence of those disorders on learning may be major. But dyspraxia does not simply consist in motor disorders or even sensory-motor coordination disorders: purely sensory disorders may also be incorporated in the concept of dyspraxia and may be more or less marked and affect to variable degrees the principal sensory systems involved in motoricity: the proprioceptive, visual and vestibular systems. In fact, all actions involving gestures and/or posture also require intact perceptual, visual and spatial capabilities, so that it is difficult to dissociate pure praxis phenomena from those related to perceptual capability.

There may be a ‘comorbid basis’ for developmental coordination disorders (DCD), attention deficit/hyperactivity disorder (ADHA), reading disorders and spoken language impairment (SLI). The question of the links between coordination disorders and reading difficulties is the subject of particular interest. One of the major questions that have yet to be resolved is whether there is a particular form of dyslexia which specifically accompanies sensory-motor coordination disorders in subjects with dyspraxia. In most of the studies on the subject, over half the subjects with dyspraxia were also found to present with reading disorders. The coincidence between motor disorders and learning disabilities in general and reading in particular is one of the bases of the cerebellar theory, which constitutes one orientation in researchers’ current quest to further elucidate the neurocognitive deficits underlying learning disabilities.

Clinical experience shows that it is rare for a child with dyspraxia to have intact writing skills. In contrast, many subjects with dysgraphia do not show clear signs of dyspraxia. The links between DCD and learning disabilities may reflect two scenarios: either dyspraxia with multiple and variable consequences on actions including writing or a more circumscribed disorder affecting writing itself and then closely linked to performances in the linguistic area.

Statistical studies of the association between precociousness and learning disabilities are rare. Practically no scientific publication has been devoted to the association between precociousness and learning disabilities, although, paradoxically, educational commentaries in non-scientific form are abundant. There are tens of articles or reports available on the Web. The lowest estimates indicate that 2 to 10% of the children enrolled in American studies on gifted children also suffer from specific learning disabilities. This is not significantly more frequent than in the overall population, but about 40% of those gifted school children do not undergo diagnosis before high school. Incontestably, the literature remains insufficiently precise in the field, preventing determination of the exact frequency of the association between precociousness and learning disabilities. Further studies are necessary. If the finding is confirmed, however, considering intellectual precociousness as a disorder concomitant with dyslexia and other learning disabilities would be of undeniable theoretical importance and constitute an important instrument for tackling the question of the underlying mechanisms. In that case, it would be necessary to image that a common process is capable of inducing both dysfunction of one module and enhanced function of the other. This places a considerable restraint on the potential explicatory models.

Associated behavioral or emotional disorders

Several published studies provide significant evidence of a frequent association between learning disabilities and behavioral or emotional disorders. Emotional or behavioral disorders may then be secondary to failure at school and/or the social, familial or affective environmental conditions exacerbating the cognitive disorder, or, on the contrary, constitute a real concomitant disorder with a link, perhaps of a genetic nature, between the two diseases. Management of those two types of disorder is possible even though the question of the mechanism underlying their association has yet to be elucidated.

Large-scale child cohort studies using a methodology based on standardized questionnaires such as the Child Behavior Questionnaire (CBQ) show that children presenting with psychological disorders (externalized or internalized) have a reading and/or mathematical level lower than children free from those disorders. The studies do not describe the cognitive profile of the children or the severity indices of the learning disabilities. They do not enable recognition of a child population that, due to poorly managed dyslexia and dysorthography and failing at school, presents with secondary psychological disorders.

Anxiety, depressive mood, attention deficit/hyperactivity disorder (ADHD) and behavioral disorders are frequently observed in children who read poorly. The association was evidenced as of the first evaluation in longitudinal studies over 7 to 10 years. The possibility of early treatment may be envisaged. The association exists independently of the other familial and social environmental factors that may be causal in emotional and behavioral disorders.

Behavioral disorders are more frequent in dyslexic children or children presenting with dyscalculia than in children free from learning disabilities, but less frequent than in the child population referred for psychiatric treatment. Performance anxiety, difficulties in peer relations, family conflicts and low self-esteem are frequently encountered in that population. By way of an example, a French reference center for language disorders found, in the course of psychological and speech-therapy medical examinations, affective disorders in 28% of 173 children presenting with a learning disorder. This calls attention to the fact that the clinician must be attentive to the cognitive and affective or behavioral aspects in the diagnostic evaluation and management of learning disabilities.

Abundant literature on the association between learning disabilities (particularly dyslexia) and attention deficit with or without hyperactivity disorder is available. All the studies agree on the fact that a child with dyslexia or attention deficit/hyperactivity disorder (ADHD) has a high risk of having another disorder. However, the selection of measurement instruments, criteria defining ADHD and learning disabilities, child age and recruitment location may explain the disparity in the frequency of comorbidity. Several causality hypotheses have been advanced and the question of the mechanism underlying the association is far from being resolved. For certain authors, the behavioral disorders may be a consequence of failing at school, or, on the contrary, for others, the reading difficulties of hyperactive children may be explained by the attention deficit. For other authors, there is a link, doubtless of a genetic nature, between the two diseases as has been suggested by the study of twins conducted by the Colorado Learning Disabilities Research Center. The higher frequency of behavioral disorders in dyslexic subjects seems directly related to the comorbidity between dyslexia and hyperactivity and to be specific to boys, while the association between anxiety or mood disorders, on the one hand, and dyslexia, on the other, is independent of hyperactivity and more frequent in girls.

It would appear indispensable to evaluate the attentional ability and degree of hyperactivity together with the other associated psychopathological symptoms in children presenting with a learning disorder in order to enable complete management.

Several teams have reported their experience as pediatric psychiatrists, psychologists or psychoanalytically-oriented psychologists with regard to children presenting with learning disabilities. When requested by teachers or rehabilitators, the psychiatrists and psychoanalysts addressed the psychological functioning that provided the context for the child’s learning difficulties in order to better elucidate the therapeutic orientations to be proposed. Studies using projective tests in dyslexic child populations do not show a univocal personality organization in those children, compared to dyslexia-free controls, excluding a unique causal relationship between personality and dyslexia. Nonetheless, when the child and his/her family are in distress and/or when rehabilitation does not yield the desired results, analysis of the child’s psychological development and interactions with his/her environment using psychoanalytical approaches, may constitute a valuable complement to the cognitive approach. Providing the two approaches are not mutually exclusive, they enable management of the whole child in his/her diversity.

Theories explaining dyslexia: phonological theory

Among the hypotheses advanced to explain dyslexia, the phonological theory is well supported by empirical studies.

The basis of the phonological theory consists in the fact that reading is a language activity. The theory is based on the finding that written language, which arose after spoken language in phylogenesis (the development of the human species) is also set up after spoken language in ontogenesis (individual development). Moreover, irrespective of the writing system, the written language is, intrinsically, a second system relative to the spoken language: even though perceiving writing depends on sight while perceiving the spoken language depends on hearing, the reader can always access the spoken form of the words he/she reads. Lastly, the phonological theory is based on the definition of dyslexia, which is a specific disorder of learning to read and does not result from patent sensory deficiencies.

Irrespective of the writing system, the word is the basic unit of writing. This explains why specific reading skills are based at the level of the procedures identifying written words. The development of automatism with respect to that identification enables the child to achieve a degree of reading comprehension equivalent to that of oral comprehension by freeing the child from slow and cumbersome decoding. In alphabetical writing, the identification may be obtained by decoding, based on the grapheme-phoneme correspondances, or by the lexical procedure which is based on the words. However, decoding does not merely consist in the beginners’ laborious reading: the expert reader may identify words that he does not know in a few hundreds of milliseconds. Moreover, the lexical procedure is, which is not based on the silhouette of written words (their overall form), enables the expert to achieve access in a few hundreds of milliseconds, not only to the visual code but also to the phonological and semantic code.

The phonological hypothesis is based on the finding that the reading performances of dyslexic subjects are notoriously weak when those subjects cannot rely on their lexical knowledge, i.e. when they have to decode rare words or non-existent words (pseudo-words). All the studies of undifferentiated groups of dyslexic subjects have evidenced that deficiency, including relative to younger children with the same reading level. Moreover, the deficiency, which is more marked when dyslexic subjects are confronted with opaque writing, is mainly observed in the slowness of the response when the writing is transparent. Lastly, in most of the cases, the phonological reading skills of dyslexic subjects are inferior to those of younger children with the same reading level. The results suggest that deficiency in the phonological reading procedure is robust and prevalent.

The conventional phonological hypothesis explains the reading difficulties of dyslexic subjects by the weakness of their phonological skills outside of reading and including in phoneme analysis and phonological short-term memory. Those deficiencies may interfere with decoding given that, in order to use the procedure, the graphemes must be matched with the phonemes. This calls for phoneme analysis capability. The units resulting from decoding have then to be assembled in order to access the words. This requires solicitation of the short-term phonological memory. More recently, deficiencies in access to the spoken lexicon were reported for subjects with dyslexia. On the basis of that finding, some researchers assumed that the origin of reading deficiencies in dyslexic subjects is dual: one origin is related to analytical skills and phonological memory while the other is related to lexicon access, generally evaluated by the response time in tasks of fast naming of images of objects or colors. This hypothesis is supported by the fact that succeeding in that type of task explains a unique part of the variance in reading in addition to that explained by phonological capability. Moreover, phonological capability and the ability to rapidly name are not correlated with the same reading skills. The former explain response precision while the latter explain processing time. Those results may nonetheless be due to the type of measurement used (accuracy scores for the phonological tasks, processing time for the others) and not to the type of task. It is currently admitted that fast naming tasks evaluate phonological skills which, when the child mainly uses decoding, enable him/her to access the spoken word matching the sequence of letters decoded rapidly and precisely.

The conventional phonological hypothesis is robust: in most of the studies conducted on groups, phonological deficiencies were found in the dyslexic subjects. In studies of individual data, a phonological deficiency was identified in most of the cases of dyslexia. Lastly, the phonological capabilities are the only capabilities that explain the reading performances of subjects with dyslexia with regard to both concomitant performance (as evidenced by regression analyses) and future performance (as evidenced by longitudinal studies). In addition, neuronal dysfunctions have mainly been detected in the brain areas involved in language processing (left perisylvian area). The dysfunctions mainly consist in hypo-activation of the areas dedicated to phonological processing, which is a strong argument in favor of the hypothesis. However, in those studies, the ‘phonological’ factor incorporates various capabilities. Deficiencies in those various fields could, in fact, be explained by an underlying factor: either a deficiency in rapid auditory processing or a deficiency in phoneme discrimination.

The auditory hypothesis stipulates that the phonological deficiency in dyslexia derives from deficient auditory perception affecting the processing of short sounds and fast temporal transitions, irrespective of whether or not the sounds of language are involved. In the current state of the research, the auditory disorders, when they are present, do not seem related to the rapidity of stimulus succession or to the order of appearance of the stimuli. In addition, compared to the conventional phonological deficiencies, the deficiencies are not frequent. Lastly, they do not enable explanation of the deficit in reading skills in dyslexic subjects when correlation and regression analyses are conducted.

Under the terms of another hypothesis, dyslexic subjects have difficulties discriminating between phonemes. This explanation is based on the fact that in order to map graphemes to phonemes, phonemes must not only be isolated but also discriminated. While numerous studies have addressed the links between learning to read and initial phonemic segmentation capability (as evaluated by counting or phoneme deletion tasks), few studies have addressed the incidence of the quality of phoneme representations on learning. However, the phoneme is the result of an arbitrary and specific slicing of language. In an acoustic continuum, we categorize, which means that we perceive a range of acoustically different sounds as /p/ and others as /t/ or /k/. Secondly, the phoneme repertory differs between languages. The phoneme is the minimal unit enabling differentiation of two words. Thus, /b/ and /v/ are two different phonemes in French which enable differentiation of ‘bol’ and ‘vol’, but not in Spanish. In contrast, the simple /r/ and rolled /r/ are two different phonemes in Spanish that enable distinction between ‘pero’ (‘but’) and ‘perro’ (‘dog’) while, in French, the two forms of /r/ are only dialect variants, allophones, of one phoneme. Neonates have been shown to perceive different phonetic oppositions liable to intervene in the world’s languages. As a function of the neonate’s linguistic environment, the repertory will very rapidly be restricted to the phoneme categories of value in processing the infant’s language. This calls for a process of selection and restructuring of the initial phonetic categories. This process is considered not to have been completed in future dyslexic subjects or at least in some of them.

The studies available indicate that in dyslexic subjects the phoneme categories are not as clearly specified as in normal readers. Dyslexic subjects discriminate phonemes belonging to different phoneme categories in their language less well (for instance, /b/ and /d/). In contrast, they have better perception of certain phonetic differences which are not of value in processing their language (allophonic perception). This double deficiency, which is not related to an auditory deficit, may only have minor consequences on spoken language acquisition. Access to the mental lexicon may take place using allophonic representations even though it may be less economical than if it were based on phoneme representations, at least in terms of the volume of information to be processed. In contrast, that type of deficiency may seriously interfere with the acquisition of written language: in order to match graphemes with the corresponding phonemes, clearly specified phoneme categories are necessary. These results may explain the difficulties of phoneme analysis, but also those of phonological short-term memory due to an increase in the memory load related to an expanded phonological repertory, that is allophonic rather than phoneme-based. The results may also explain the difficulties in accessing the lexicon, in particular those highlighted by serial rapid naming tasks. Hence a syndrome underlying the various deficiencies incorporated in the classic phonological explanatory framework for dyslexia may exist.

The results obtained with phoneme discrimination tasks provide further arguments in support of the explanatory hypothesis of phonological deficiency in dyslexia. However, complementary studies are necessary in order to evaluate, first, the frequency of the deficits, and, secondly, their explanatory power and the relationships they maintain with conventional phonological deficiencies (among others, phoneme segmentation and phonological short-term memory deficiencies).

The visual dimensions of dyslexia

Another explanatory hypothesis for dyslexia covers the visual dimensions. Learning to read requires matching orthographic information derived from visual analysis of the sequence of the written word and phonological information derived from auditory processing of the corresponding spoken sequence. Numerous studies have addressed the nature of the visual processing involved in that learning and have formulated the hypothesis that certain dysfunctions in visual processing could induce dyslexic disorders.

Work conducted on that context suggests that certain subjects with dyslexia have, in fact, a deficiency in visual processing (independently of any perceptual impairment). It is important to note, however, that those visual disorders have never been evidenced by clinical tests and, in order to evidence them, computerized psychophysical tests are required.

The results of a number of empirical and neurophysiologic studies argue in favor of impairment of the magnocellular visual system in subjects (adults and children) presenting with developmental dyslexia. This conclusion was based on the fact that certain studies evidenced a disorder of sensitivity at low spatial frequencies and high temporal frequencies together with reduced sensitivity to moving points: the performance of dyslexic subjects is inferior to that of normal reading subjects with regard to detecting movement of a set of points. Dyslexics are less able to discriminate the difference in speeds of two moving targets. Some studies have also shown the existence of correlations between the performances obtained in psychophysical visual tests and various reading measurements. In addition, the magnocellular hypothesis, initially restricted to the visual system, has gradually evolved to tend toward the hypothesis of an amodal disorder of the magnocellular systems: dyslexic children are reported to have difficulty processing fast temporal information in both modalities (visual and auditory) subsequent to joint impairment of the visual and auditory magnocellular systems. The results of empirical and neurophysiologic studies would appear to support the hypothesis and thus account for the concomitant occurrence of phonological and visual disorders in certain dyslexic subjects.

However, the magnocellular hypothesis is currently controversial. The methodological limits of certain studies that found a magnocellular visual deficiency have been stressed and the hypothesis has suffered from lack of result duplication. The amodal hypothesis is also to be reviewed in the light of results suggesting that only a small proportion of dyslexic subjects with a phonological disorder also have difficulty in processing the sounds of speech and difficulties in magnocellular visual processing. However, the research conducted in that context points to the extreme complexity of the type of disorder in question. It may now be considered that the disorder is manifested under certain particular experimental conditions that have largely still to be defined and is only observed in a subpopulation of dyslexic children that has also not been clearly defined.

The most recent studies suggest, in particular, that the magnocellular disorder could only become patent when the task involves specific attentional processing. This is in line with the results of a certain number of other studies showing the existence of visual attention disorders in contexts of dyslexia. Difficulties in attention focusing, attention disengagement and problems of automatic orientation of the attention giving rise to a left mini-negligence phenomenon have been described in dyslexic subjects. However, the hypothesis of a visual attention disorder like the hypothesis of magnocellular impairment is confronted with the fact that the deficiencies have most frequently been observed in association with phonological disorders. Thus, visual attention disorders are to be viewed in the context of an amodal disorder of attention processing extending to the auditory modalities and perhaps also haptic modalities in addition to the visual modality. It is, however, important to note that the deficiencies evidenced in dyslexic subjects are not situated in the context of a general attention disorder (e.g. ADHD). Few studies have provided any evidence elucidating the link between visual attention disorder and the reading profiles of dyslexic children.

Most frequently, the visual or visual attention dysfunctions have been described in the context of phonological disorders. Recent studies suggest, however, that a particular form of dysfunction, a disorder of the visual and attentional span, may be associated with certain forms of dyslexia and be observed independently of any phonological impairment. The concept of the visual and attentional span refers to the number of letters that can be simultaneously processed in the context of a word letter sequence at each glance. A reduction in the visual and attentional span has been evidenced in certain cases of surface dyslexia with no associated phonological disorder. Group studies suggest that the number of children presenting with that type of deficit is at least equal to the number of children presenting with an isolated phonological disorder in both the English- and French-speaking populations. Studies need to be continued in order to test the hypothesis that there is a causal relationship between visual and attentional span deficiency and difficulty in learning to read.

All the professionals agree that a diagnosis of dyslexia can only be formulated after having checked that no visual perception disorder is present: an ophthalmologic examination is therefore necessary in order to determine the child’s visual acuity and rule out problems such as hypermetropia, myopia or astigmatism. It is also necessary to question the child on the feelings experienced when reading. Some dyslexic children have the impression that the letters move or overlap during reading. This reflects instability of binocular control. Any complaints of that type together with a number of warning signs (visual errors, difficulty following the lines, jumping from line to line) require complementary investigations (orthoptic examination and evaluation of the child’s binocular focusing). The clinical examination must also rule out oculomotor disorders (e.g. nystagmus or exophoria). Management methods such as colored lenses, prisms or temporary occlusion of one eye are not unanimously approved either by researchers or ophthalmologists. It must be noted that those ‘treatments’ have yet to be firmly validated and a convincing theoretical context has yet to be formulated. Instruments for the clinical diagnosis of magnocellular or visual and attentional disorders are lacking. Research needs to be pursued in order to develop the training necessary for management of those disorders.

Cerebellar hypothesis

Again in an attempt to account for the complexity of the clinical symptoms observed and, in particular, the frequent association, within learning disabilities, of deficits in fields as varied as reading, language, arithmetic and even the sensory-motor systems, scientists have searched for explanation and proposed models tending to extract the shared characteristics of the various disturbed areas.

Thus, certain associations observed in dyslexic subjects have attracted the attention of researchers: a delay in the stages of motor development, sequential and temporal disorders (say the time, remember the month of the year) and, above all, the presence of motor coordination and balance disorders, all suggest dysfunction of the cerebellum. The latter organ, which was for many years considered to play a purely motor role, has been the subject of recent studies that have demonstrated its involvement in numerous cognitive processes and in learning in general.

Empirical evidence has supported a theory that was essentially based on clinical intuition: firstly, anatomic studies of the human brain post mortem and using various imaging methods have detected an anomaly of the cerebellum in dyslexic subjects.

Rather paradoxically, a study demonstrating cerebellar hypo-activation during purely motor tasks (learning a series of finger movements) in dyslexic adults was to draw attention to the cerebellar theory. In contrast, although the cerebellum is classically activated when a normal reader reads, almost no reports of cerebellar activation deficit when a dyslexic subject reads have been published.

Hence, the cerebellum may affect reading in different ways. The cerebellum is involved in control of eye movements, visual and spatial attention and peripheral vision, all essential components of reading. As a crucial structure in time management by the brain, the cerebellum may contribute to the sensory-motor coordination and intense integration problems observed in dyslexic subjects.

According to the advocates of the cerebellar theory, it supposes that the deficiency is present very early, as of birth, and interferes with the normal setup of auditory and articulatory skills necessary for constitution of the phonological system, including visual abilities such as eye movement and letter recognition, giving rise to both the phonological and orthographic difficulties characteristic of dyslexic children and adults.

As attractive as the theory may seem, it has been greatly criticized in recent years. The critics observe, first of all, that motor disorders are far from being the rule in all dyslexic patients and that many dyslexic subjects, even those with major difficulties in learning to read, show no motor difficulty or indeed are particularly gifted for motor activities both proximal and balance motoricity and distal motoricity. Certain authors have even suggested that the motor disorders sometimes observed in dyslexic subjects are only an artifact related to the coexistence of hyperactivity disorders. While several studies have thus minimized the incidence of such motor disorders, at least two recent studies, using sophisticated balance and posture measurement systems, have contributed positive arguments in support of the cerebellar hypothesis. Dyslexic subjects were found to be significantly less able than controls to maintain balance on one foot, in particular with the eyes open. The performance of the dyslexic subjects in the postural test correlated with their reading and orthographic performance. Lastly, if, as has been reported, cerebellar-type disorders are less frequent in dyslexic adults than in dyslexic children, this perhaps means that the deficiencies initially observed in children stabilize over adolescence and have resolved by the time the dyslexic subject has become adult. In short, it is improbable that dyslexia can be explained by cerebellar dysfunction. However, the arguments are sufficiently numerous to encourage inclusion of the cerebellum among the brain systems, which are disturbed in at least some dyslexic subjects. In any event, the cerebellar hypothesis has had the merit of proposing a plausible alternative to the theses in which the reading disorder was the sole subject of interest to researchers. The cerebellar hypothesis has thus opened the way to taking into account comorbidity in the explanation of dyslexia.

The temporal processing hypothesis

Not unrelated to the cerebellar theory, a hypothesis was proposed over 30 years ago by an American scientist, Paula Tallal, under the name: ‘Temporal processing deficit theory.’ The dyslexic child and, more generally, children suffering from specific learning disabilities, frequently have problems with time in general irrespective of whether they consist in problems with management of the temporal aspects of everyday life, consciousness and/or the perception of the duration of events or discrimination between brief events such as the events constituting human speech. Very many studies using a variety of approaches have addressed that strange characteristic with the aim of identifying a common point between the latter and learning difficulties.

According to the theory initially proposed by Tallal, the brain of a child suffering from spoken and written language disorders is constitutionally unable to specifically process the environmental stimuli characterized by briefness and rapid succession which are precisely the two characteristics of human speech. This finding becomes meaningful when we observe the auditory confusions to which many dyslexic children are subject. Dyslexic children tend to confuse acoustically close phonemes, in particular pairs such as /t/-/d/ and /ch/-/j/ which seem particularly difficult to distinguish for the auditory system of a dyslexic subject. Among the tests enabling evidencing of temporal auditory processing deficit, Tallal’s ‘repetition test’ is doubtless the most widely used. The subject is presented with pairs of different sounds in random order. The child is to reproduce the sounds by successive pressures on two keys of a computer keyboard. By varying the interval between the two sounds from a few milliseconds up to several seconds, it is observed that the dyslexic subject experiences marked difficulties in reproducing the order but only for short intervals of less than 150 milliseconds. For longer intervals, the dyslexic child’s performance becomes similar to that of the non-dyslexic controls. This disorder in temporal sequencing is found in children presenting with language disorders in general but also in at least some of the children presenting with dyslexia. In addition, the same effect can be demonstrated in dyslexic subjects by using pairs of syllables such as /ba/-/pa/ or /da/-/ga/. The simple test gave rise to a vast debate. The defenders of the temporal hypothesis consider it to demonstrate that dyslexic subjects suffer from a difficulty that is both perceptual and temporal, while the critics stress that the effect is not present in all dyslexic subjects and, above all, varies depending on the linguistic or non-linguistic nature of the stimuli. If children are required to fulfill a temporal order judgment (TOJ) task using pairs of phonemes that are easier to discriminate, the deficit disappears. The same applies if the stimuli consist in artificial speech.

Another important consideration is related to the potential heterogeneity of the concepts based on the term ‘temporal deficit’. While TOJ has been most widely studied and, in general, found to be deficient in dyslexic subjects, it has rarely been related to other aspects of temporal processing such as the judgment of the relative duration of a stimulus. In addition, it is important to consider the links between a potential temporal disorder and the intensity of the phonological disorder considered the crucial mechanism in learning to read. A recent study has shown that TOJ deficit is significantly correlated with the phonological disorder and a measurement of the judgment of auditory stimulus duration. However, the two deficits were not inter-correlated suggesting that they consist in two separable dimensions of the temporal deficit.

Another postulate of the temporal theory predicts that the deficit is to be found in different modalities. Several studies, including one by Tallal’s team, have demonstrated that dyslexic subjects have difficulty discriminating between two stimuli in tactile mode (identifying which of two fingers of the same hand were touched simultaneously). More recently, various studies have evidenced judgment deficits with respect to visual and tactile stimuli in dyslexic subjects, including both in the same subject. This supports the idea of a supra-modal disorder, i.e. a disorder independent of the sensory modality. Lastly, several authors have recently stressed a particular difficulty in dyslexic subjects with regard to discriminating between stimuli not only in several perceptual systems in a given subject, but in tasks involving confrontation of several modalities in a given task, thus implementing an inter-modality transfer.

In both dyslexic children and adults, the protocols evidenced very marked differences for the majority of combinations studied. For the authors, these results show that, beyond the multimodality impairment suspected by Tallal as of the initial formulations of her theory, there also exists a deficit in the temporal processing necessary for the pooling of information reaching the brain via different sensory channels in the dyslexic subjects. A recent study using a visual temporal order judgment paradigm associated with an auditory facilitator, suggested that dyslexic subjects have a problem with respect to the width of the ‘temporal window’. The abnormal width of the temporal window in dyslexic subjects is considered to impair the process depending on rapid and precise pairing of two items of information derived from different modalities such as those involved in grapheme-phoneme conversion.

Thus, after a period in which the theories based on a supra-modal temporal disorder were sharply questioned, the more recent research, using refined study protocols, seems once again to argue in favor of a deficiency in certain very specific aspects of sensory information processing, in particular when two different items of information are to be related. The temporal characteristics of relating those items may be determinant. These findings are reminiscent of an already old theory of learning known as ‘Hebb’s synapse’: when two neurons, A and B, are in proximity and neuron A discharges when B is activated, the reciprocal links between A and B are strengthened. Conversely, if A discharges when B is inactive, even if the discharge takes place within a few milliseconds, the links between the two neurons are inhibited. The general principle, which since it was first formulated in the 1940s has been confirmed empirically several times, could well prove, on the basis of the most recent data, a propitious framework for explaining the diverse and still poorly elucidated findings with regard to learning disabilities.

With regard to the auditory modality, a large number of studies have used the evoked potential method, of particular value in this context because of its temporal sensitivity: various studies have shown that the auditory cortex of the dyslexic subject imperfectly processes auditory stimuli with, specifically, at least in certain cases, evidence of an anomaly in the processing of the temporal succession of the acoustic components of linguistic units. Thus, the acoustic difference between the phonemes /ba/ and /pa/ (the concept of voicing) is reflected in the electrical activity recorded opposite the auditory cortex by a subtle difference in the temporal succession of events strictly aligned temporally with the acoustic events. In dyslexic adults, the same stimulus is processed in a temporally anarchic manner such that the usual succession of the acoustic components is no longer recognized at electrical level. It may be presumed that a minimal disturbance in the simultaneity of activation of the various cell components of the system, by reducing the strength of their reciprocal connections, would be sufficient to prevent acquisition of specific perception of voicing simply because the latter is, among the characteristics of the human voice, the most dependent on the temporal organization of the stimuli.

The contribution of brain imaging

The neurological approach to dyslexia in contemporary research began with the founding histological study by Galaburda et al. who showed the presence of micro-structural anomalies in the perisylvian cortex (ectopias and dysplasias) and a reduction in the size of the neurons in the geniculate nucleus. Those results suggest the existence of anomalies in neuronal migration and maturation in certain zones of the cortex essential for the automatism of reading mechanisms.

Given that basis, morphometric studies using magnetic resonance imaging (MRI) were conducted to detect the existence of macroscopic anomalies in cortical structure in dyslexic subjects. Targeting predefined regions of interest (parietal cortex, temporal cortex, inferior frontal cortex, cerebellum, corpus callosum), the studies detected structural anomalies but with no great degree of between-study reproducibility. Technological progress enabled analysis of the whole brain volume using voxel-based morphometry, a diffusion-tensor imaging method. The results obtained with the latter method showed particularities of the signal generated by a zone of white matter underlying the left temporoparietal junction suggesting abnormal connectivity between those zones of the cortex. In addition, correlations between anomaly intensity and reading score were observed.

Cerebral functional imaging (positron-emission tomography (PET) and functional magnetic-resonance imaging (fMRI)) and electromagnetic techniques (electroencephalographic-evoked potentials (EP), magnetoencephalography (MEG)) enabled detection of anomalies in the activity of the regions involved in reading and the associated cognitive processes in dyslexic subjects vs. normal readers. It should be noted however that the studies are subject to numerous experimental biases which may explain the discrepancies between the studies.

In the field of electrophysiology, anomalies have been reported in both the early and late potentials as a function of the experimental paradigms and visual or auditory modalities investigated.

Under PET or fMRI imaging, the normal functional anatomy of reading isolated words involves three major areas: the left temporo-occipital junction, the left temporoparietal junction and Broca’s area. Activation of the 3 regions is abnormal in dyslexic subjects as is the functional connectivity between those regions. The two temporal and/or parietal regions are very generally hypo-functional. This has been confirmed by MEG studies. Hyperactivation of Broca’s area in patients has frequently been considered to reflect premotor compensation for posterior cortex deficiency. Compensatory effects have also been suggested in order to explain the more marked activation of the right perisylvian cortex in dyslexic subjects.

The results in adults have broadly been confirmed by fMRI and MEG studies of children. Overall, the studies suggest an anomaly of connectivity with respect to the temporoparietal and frontal circuits which underlie language and particularly the circuits subjacent to the phonological loop in working memory and the interaction between the ‘dorsal and ventral reading circuits’.

Numerous pathophysiological hypotheses with regard to the etiology of dyslexia have all been empirically supported by brain imaging studies confirming the existence of the predicted differential effects. The hypotheses may be divided into two types. The first type postulates the existence of a deficit affecting the phonological or lexical representations stored in long-term memory (phonological hypothesis). The second type covers numerous hypotheses suggesting a variety of anomalies in the sensory-motor processes for real-time processing of perceptual information.

The phonological hypothesis, which is predominant, has been greatly strengthened by the existence of activation deficits in the left perisylvian regions normally involved in phonological analysis and phonological working memory.

Certain studies have suggested a dysfunction of the magnocellular visual pathways but others have shown that those effects were only evidenced under certain empirical constraints. In the auditory field, several studies have shown a deficit in processing, of the magnocellular type, with respect to rapid acoustic or linguistic variations. However, recent data suggest that those effects are only present in subgroups of subjects and only involve certain sectors (left premotor) of the cortex while other sectors (left supramarginal) show constant hypo-activation irrespective of the variation rate of the speech signal.

Lastly, anomalies of right cerebellar cortex activation were observed during tasks consisting in learning a sequence of finger movements.

The neurological imaging methods were more recently applied to the study of the cerebral correlates of the improvement in performance induced by various interventional or intensive training methods for groups of dyslexic children. The greater influence on performance and neurofunctional signals of intensive phonological intervention, compared to conventional management, was demonstrated in some of those studies. In other studies, positive correlations between the increase in the fMRI activation signal generated by the left temporoparietal and right frontal regions and improvements in language and phonology scores, without generalization to the reading score, were observed. The various methods of functional imaging are undoubtedly sufficiently sensitive to evidence changes in cerebral activity following remediation but the specificity of the mechanisms underlying the effects obtained at cognitive and cerebral level are to be discussed.

So far, the functional or anatomical cerebral imaging studies have shown neurological correlates with the cognitive disorders. However, the direction of the causal relationship could be determined by longitudinal studies initiated in young children. Those studies would enable the effects of cerebral anomalies present from the outset to be distinguished from those resulting from adaptation to the handicap.

Lastly, functional cerebral imaging could play a role in the very early identification of the cerebral activity profile enabling prediction of the emergence of a handicap at learning to read level, particularly in families at genetic ‘risk’ of developing the disorder. An ethical assessment of that research orientation is to be conducted due to the risk of stigmatization when the resources necessary for handicap management are not ensured.

Genetic factors

The hypothesis of the familial nature of dyslexia was first suggested long ago and circumstantial evidence collected over 20 years does indeed point to a genetic origin of dyslexia.

The best indicator of a familial component is an increase in risk in the relatives of an affected subject. The risk is a relative risk comparing the risk of relatives of healthy subjects to the risk of the disorder in the relatives of subjects suffering from it. In general, the first-degree relatives’ (parents, brothers and sisters) history is used. For dyslexia, the risk of the disorder in a relative of an affected subject is 8-to 10-fold higher than the same risk for the relatives of an unaffected subject. This does not exclude the occurrence of other sporadic cases insofar as the same cerebral and cognitive dysfunctions may occur de novo. Family clustering suggests but does not prove a genetic origin. Families share not only some of their genes but also a certain environment. It may be considered that parents who do not read constitute an environment that is not conducive to learning to read for their children.

Studies of twins enable optimum assessment of the weight of the genetic factors compared to that of environmental factors. The studies are designed to calculate the inheritability, in other words the percentage of the variance explained by genetic factors. In order to do so, the concordance between the impairments of monozygotic and dizygotic twins is measured. When a monozygotic twin is dyslexic, the probability of the other twin also being dyslexic is about 70%. In contrast, the probability is only 45% for dizygotic twins. Since monozygotic twins may be considered to share the same environmental factors as dizygotic twins, the difference in concordance is explained by the fact that monozygotic twins are 100% genetically similar while dizygotic twins are only 50% similar (for the genes which vary). The concordance data thus enable calculation of the inheritability. This does not, however, indicate a direct causal relationship. A change in the environment of two true twins during their intrauterine development or during the perinatal period (lack of oxygen, maternal hormone, toxic product, etc.) may induce the same effect after birth through similar responses to an environmental change. The hypotheses deserved to be studied in the field of dyslexia.

The inheritability of dyslexia is between 50 and 65% according to the larger scale studies on twins. Similar data were collected with regard to mathematical difficulties with an inheritability of about 50%. With regard to dysorthography, the genetic studies did not truly distinguish it from dyslexia. Many studies used both orthographic and reading measurements to define the phenotype. For the time being, no study has specifically investigated for genetic factors which might distinguish dyslexia from dysorthography.

The data show that there is indeed a genetic contribution to specific learning disabilities but do not identify the genetic factors or explain their mode of action. That is the role of molecular and neurogenetic studies. At present, such studies have only begun generating results for dyslexia. With regard to the other specific learning disabilities, molecular genetic research is only at its inception.

In the case of dyslexia, numerous analyses of the genetic link have shown chromosomal regions that are significantly more frequently transmitted from the parents to children with disorders than to children without disorders. The link between dyslexia and regions on chromosomes 1, 2, 3, 6, 15 and 18 has been reported by independent teams on the basis of several studies. The multiplicity of the chromosomal sites identified suggests that the specific learning disabilities are, in most cases, deficiencies with a complex genetic component in which several genes are involved.

The linked chromosomal regions indicate regions of the genome in which the genes associated with dyslexia are very probably located. However, the regions may contain hundreds of genes that have been inventoried but that, in the majority of cases, have never been studied. There is thus an enormous gap between identifying a linked region and identifying the associated gene in that region. Progress in molecular genetics is reducing the magnitude of the gap day by day.

In the six regions, four candidate genes have been proposed. Two of them (DYX1C1, ROBO1) show mutations in rare cases or families but it is not known whether the alleles of those genes increase the risk of the more common forms of dyslexia. For the other two genes (KIAA0319 and DCDC2), susceptibility haplotypes in large populations have been suggested. The haplotypes have nonetheless to be confirmed. Outside of the rare cases of a radical mutation impairing protein function (DYX1C1 in a Finnish family), it would appear that susceptibility alleles exert their effects through alteration of the expression of the protein (certain haplotypes are, moreover, situated in regulatory regions). Nonetheless, the expression data remain fragmentary and require confirmation.

The most remarkable result with respect to the functional properties of the genes was generated by studies on developing animals. The studies show that four of the candidate genes for dyslexia are involved in the migration of neurons during cerebral development, three in the radial migration of cell bodies towards the cortex and one (ROBO1) in axon and dendrite migration. It seems highly unlikely that the convergence of functions is random. Moreover, the results enabled a link to be established between the genes associated with dyslexia and the cerebral abnormalities observed in the brains of certain dyslexic subjects.

It is important to note that while the genetic usage is such that the genes are referred to as ‘genes for dyslexia’, this is a misuse of language and in fact a shortcut for saying ‘genes of which certain alleles increase the risk of dyslexia’. It is obvious that none of the genes is specific for dyslexia or is a gene for reading or even a gene for spoken language. Like almost all human genes, related forms are found in other mammals or even in Drosophila (ROBO1). In all animals the genes exert multiple functions and are expressed in numerous different organs during development and life. What links the genes more specifically to dyslexia is the fact that they are also involved in a particular stage of brain development and notably in the setup of certain brain areas which are later recruited for learning to read.

In certain rare cases, it would appear that mutation of a single gene is sufficient to induce dyslexia. The gene responsible may vary between cases. Most frequently, dyslexic subjects do not seem to carry a rare mutation but rather susceptibility alleles. The alleles are frequent in the normal population and do not in themselves constitute a sufficient cause of dyslexia. Each susceptibility allele slightly increases the risk of dyslexia. The disorder only emerges when particularly unfavorable combinations of susceptibility genes and/or interactions of the alleles with non-genetic factors combine in increasing the risk. A large number of environmental factors (biochemical, traumatic, linguistic, socio-educational, pedagogic) may modulate the expression of genetic factors in a positive or negative direction.

Thus, it is highly probable that in a certain number of cases the primary cause resides in non-genetic factors, for example factors inducing brain anomalies at birth.

Each dyslexic individual thus has his/her own cognitive profile and cerebral particularities, which are the result of the specific combination of genetic factors carried and the non-genetic factors to which the individual is exposed.

Critical analysis of the various theories of dyslexia

There is great diversity in the theories advanced to explain dyslexia: phonological deficiency theory, temporal auditory processing theory, visual theories, cerebellar theory, magnocellular theory, etc. The diversity is due to several factors:

• there are undoubtedly several distinct causes of dyslexia and thus more than one theory could be correct, each theory accounting for a subset of the dyslexic population;
• the clinical presentation of dyslexia is complex and includes numerous symptoms other than reading, particularly phonological, auditory, visual, spatial, motor and other symptoms. Each of the symptoms has given rise to theoretical speculations;
• dyslexia is frequently a concomitant disorder with other developmental disorders (spoken language disorder or dysphasia, dyspraxia, attention disorders, etc.). In the studies of groups, the symptoms of other developmental disorders may thus seem linked to dyslexia giving rise to new theoretical hypotheses.

The wide variety of symptoms associated with dyslexia does not facilitate identification of the real causes and distinction of those causes from simple concomitant disorders. Nonetheless, on the basis of a very large number of studies, a few points have clearly emerged.

The vast majority of dyslexic children suffer from a specific cognitive deficit with regard to the representation and processing of the sounds of speech. This is known as a phonological deficiency. The deficiency interferes with learning, mastery and automation of the use of the grapheme-phoneme matches and, subsequently, with all of the process of learning to read, including the orthographic route. In general, it is therefore not appropriate to distinguish between phonological and surface dyslexia.

A minority of dyslexic children seem to present with visual disorders and no phonological deficiency. It would even appear that there are several subtypes of visual dyslexia. However, empirical studies in the field are for the time being insufficient and have not generated sufficiently comprehensive theories.

With regard to the neurobiological causes of cognitive deficiencies, two main types of explanation remain possible: those centered on the biological causes of the phonological deficiency that is directly responsible for reading disorders and those attempting to account for the association with sensory-motor disorders. The former postulate a specific impairment of early development of the left perisylvian area involved in language acquisition. The hypothesis was recently strengthened by genetic data. The ultimate source of dyslexia could thus reside in a certain number of susceptibility alleles on multiple genes. The alleles, acting alone or in combination, would increase the risk of interference with neuronal migration in the left perisylvian area.

In contrast, hypotheses such as the cerebellar theory or the temporal processing theory supply a more direct explanation of the associations of concomitant disorders. However, those hypotheses are currently insufficiently supported by empirical data and require new studies, in particular with regard to their ability to reflect the links between sensory deficits and reading disorder.

Moving beyond the discrepancies in the ultimate neurobiological causes, the two types of theory converge with respect to the presence of dysfunction in the development of the cerebral areas normally involved in the representation and processing of the sounds of speech (‘phonology’). The dysfunction induces a cognitive deficit whose main manifestations are weak phonological consciousness, weak short-term verbal memory and slow retrieval of phonological representations. The cognitive deficit in general has little influence on spoken language acquisition (except in the event of very marked severity or additional language disorders). In contrast, the cognitive deficit becomes patent with written language acquisition which recruits, in a particularly intense manner, the phonological skills. Lastly, the symptoms observed with respect to reading are the product of cerebral and cognitive factors and also numerous environmental factors such as the richness of the linguistic environment (which, in particular, influences vocabulary), the regularity of the writing system, the method of teaching reading and, undoubtedly, other environmental factors. The model proposed below reflects the interactions which are liable to occur between other environmental factors (injury, affective deficit, sociological factors, pedagogic factors, etc.), at various levels, and the biological, cognitive or even symptom expression factors. The model clearly shows that an exclusively biological or exclusively environmental (irrespective of its nature) determinism is not meaningful. However, further studies are required in that context. Studies would enable investigation of all the factors in a multi-scale and multi-field model (molecular bases, neuronal networks, cognitive psychology, developmental psychology, education) with a view to elucidating the learning and development mechanisms and their dysfunctions. Such studies would thus further the understanding of the links between the deficits observed at neuronal level by means of anatomical or functional imaging and the dynamic models underlying the interaction between individuals (parents-children, teacher-child). The objective of such studies would be to tend towards unifying the various partial models in a credible common synthesis ultimately enabling researchers and professionals to understand the same ‘symptoms’ using different and complementary approaches.

Diagram of the various factors involved in dyslexia

Detection, screening and diagnosis

Detection, screening and diagnosis may be distinguished in that the systems, players and instruments differ. Unlike screening and diagnosis, detection may be implemented by persons other than healthcare professionals. The teacher, in liaison with the parents, is in the best position to detect a child who has learning difficulties. At nursery school, the teacher may detect spoken language disorders, risk factors for potential subsequent difficulties in written language acquisition, difficulties in writing or difficulty learning the numerical code. As of primary school, the difficulties in learning to read, spell, do arithmetic and write may be detected. The teacher is able to clearly define the acquisitions and gaps without being able to qualify what he/she has observed. The parents and family physician may also draw attention to the difficulties observed.

Screening is, in principle, a procedure that targets a given population. Since it targets all the children in a given age group, the school is the preferred site for screening professionals who use particular instruments to detect acquisition and learning problems. Article 85 of French act No. 2002-73 dated January 17, 2002, relating to social modernization provides for organization of language disorder screening during a mandatory assessment in the 6th year. Screening is, in principle, conducted by the National Education physicians. However, screening for scholastic learning disabilities can only begin when the children have begun to learn, in other words after the start of primary school class CP (6–7 years). Screening during the obligatory examination at age 6 years is thus most frequently a screening for risk factors for specific learning disabilities. Screening for risk factors (such as spoken language disorders) may also be conducted by the Maternal and Infantile Protection (PMI) physicians during the examination of infants in nursery school classes PS (3–4 years) or MS (4–5 years). Physicians with training in the field may be requested. During schooling, the network for specialized assistance to children in difficulty (Rased), a National Education in-house structure (consisting of a psychologist and specialist teachers) contributes to individual identification and screening.

Diagnosis frequently calls for the skills of various professionals working as a pluri-disciplinary team given the complex nature of the disorders and the frequent existence of associated disorders. The open-care professional networks frequently offer that pluri-disciplinary approach coordinated by a referring physician. Reference centers have been set up in the teaching hospitals (CHU). There are about 40 reference centers distributed through France. All the centers offer multidisciplinary consultations and, as a minimum, medical, speech therapy and psychological consultation and if necessary psychomotor and neuropsychological consultation for an overall assessment.

The instruments differ depending on whether they are intended to detect, screen or diagnose specific learning disabilities.

Before the start of scholastic learning (before 6 years), the instruments are designed to identify, screen or diagnose spoken language disorders and detect signs predictive of scholastic learning disabilities. After age 6 years, the instruments are designed to identify, screen or diagnose learning disabilities (most frequently written language disorders).

The first category of instruments enables identification of subjects at risk of subsequent reading disorders. Before age 5 years, the instruments (e.g. ERTL4: test to detect language and learning disabilities) are able to detect spoken language disorders. For children aged 5 to 6 years, the instruments (e.g. BSEDS: health assessment to evaluate development for school children aged 5–6 years) investigate spoken language disorders and identify the risk factors for dyslexia. The subjects at risk thus detected will not necessarily become dyslexic. Management of the spoken language disorders (depending on severity) is in itself a means of preventing dyslexia since management acts on the risk factor.

Instruments for screening for written language acquisition disorders that can be collectively administered like Timé 2 from January in primary school class CP (6–7 years) to the end of CE1 (7–8 years), or individually administered like Odedys, calibrated on CE1 (7–8 years), or Batelem-R as of CP (6–7 years), or the BREV learning items instrument from CP (6–7 years) to CE2 (8–9 years).

A second type of instruments enables first-line neuropsychological clinical examination (e.g.: BREV, quick battery for the evaluation of cognitive functions). The instruments are of value in determining the reality of a disorder and its profile and seriousness, such as, for example, a specific spoken language disorder at age 5 years or a written language disorder at 7 1/2 years or writing disorder after 5 years. The instruments are thus used to define the complementary assessments necessary to confirm the diagnosis and the pedagogic and care operations that are necessary. They also enable the professional to assess the course of the disorder. The clinical examination is also designed to rule out a sensory disorder or other patent neurological or psychiatric impairment.

Lastly, the third type of instrument is designed to confirm a diagnosis based on the preceding instruments. The third type of instrument includes: N-EEL (new tests for the evaluation of language) and Elola 5 (battery for the evaluation of spoken language in aphasic children) for spoken language; Belec (battery for the evaluation of written language), Odedys or Evalec (battery for the diagnostic evaluation of dyslexia); and various neuropsychological batteries for attentional functions. The Weschler composite intelligence battery enables definition of the profile of the intellectual functions.

The third type of instrument, which is very specialized, is used by the professionals involved: psychological evaluation of behavioral, affective and cognitive functions by a psychologist; evaluation of spoken and written language by a speech therapist; evaluation of graphic and praxis functions by a psychomotrician or ergotherapist or neuropsychologist; evaluation of attentional and memory functions by a neuropsychologist, etc. For each of the specific learning disabilities, diagnostic instruments enable elucidation of the child’s disorder by reference to recognized neuropsychological models. For example, for written language disorders, the instruments are not only to determine reading age but also the strategies used by the child as a function of the precision and rapidity of identification of regular and irregular words and pseudo-words and the underlying cognitive skills and comprehension.

Use of the various tools refers to key ages:

• as of age 3 years for the detection of a spoken language disorder that may lead to diagnosis of a secondary disorder (deafness, behavioral or communicational disorder, or even intellectual deficit) requiring specific management and not simply management the spoken language deficiency. This prerequisite is indispensable for the diagnosis of specific disorders. Up to the age of 4 1/2 years, only specific and severe disorders (one or several seriousness criteria: unintelligibility, agrammatism or comprehension deficit) necessitate detailed speech therapist evaluation and rehabilitation as soon as the child cooperates. Specific disorders without seriousness criteria are followed up with pedagogic adaptation and parental assistance;
• at age 5 years for screening for a writing disorder and/or dyspraxia requiring complementary psychological evaluation to confirm the specific nature of the disorder and, if the disorder is specific, psychomotor or ergotherapy assessment to determine the profile and seriousness and the care indicated;
• as of age 5–6 years for screening for numerical code acquisition difficulties (name and construction of numbers, numerical chant) and, in primary school years CP-CE1 (6–8 years) to screen for difficulties in accessing numerical facts (addition then multiplication tables) and number trans-coding. Screening may result in a psychological evaluation to confirm the specific nature of the disorder and a precise evaluation of the arithmetic difficulties;
• as of primary school class CP (6–7 years), to screen for disorders of written language acquisition. If those disorders are associated with a persistent disorder of spoken language, they require speech therapist evaluation associated or not associated with psychological evaluation;
• as of the second part of primary school class CP (6–7 years), to screen for severe disorders (non-acquisition of the process of decoding, no improvement after appropriate training, etc.) also requiring complementary evaluation of the cognitive capabilities, including those with respect to language.

The diversity of the scenarios depending on the key ages and symptoms determines the importance of the first-line screening examination.

Evaluations are indispensable in order to assess the time course of the disorder six months to one year after initiation of the educational and/or care project. The follow-up evaluations include, as a minimum, the tests whose scores were deficient in the initial evaluation in order to conduct objective quantitative and qualitative comparisons and reorient the educational and care program.

An expert commission was set up (decree dated February 8, 2002) in order to compile, at national level, recommendations on the instruments for the use of professionals in the context of the Action plan for children presenting with a specific language disability. The commission’s report addresses the instruments cited above. The instruments have just been made available to professionals in the medical and childhood sectors and are available from the Ministry of Health and Solidarity website⁸. Since the findings were made public at the time of finalization of the work of this collective expert review, it was too late to incorporate the analysis in the report.

Prevention in the school environment

Prevention in the school environment could target three populations of children: children identified as being at risk of presenting difficulties in learning to read in the GS class of nursery school (5–6 years); children experiencing great difficulty in reading in the CP class (6–7 years) of primary school; dyslexic children requiring special-needs teaching to promote learning in the preserved fields.

Numerous foreign studies have evaluated the effect of training with the objective of obtaining a pedagogic first-line response in school. Training targeted children at risk of difficulties in acquiring written language (from families at genetic risk) or, more frequently, children failing to learn to read. The prevention was not targeted on dyslexic children since it was implemented prior to any diagnosis.

The scientific bases for the type of training were the knowledge generated by basic research on learning to read. All the studies targeted English-speaking populations, i.e. populations learning to read in a language even less transparent than French.

The issues involved in those studies are essential: a pedagogic response in class can be implemented for all children without discrimination; the response constitutes a preventive action and does not induce health cost in contrast to a response consisting in care. It is thus indispensable to take the results of those studies into account: What is the most pertinent training? What children are involved? What is the quality and intensity of the effects? Nonetheless, the particularity of the French language and educational system will make French studies necessary.

The value of training is related to the need to prevent children who read poorly falling behind their peers. The child must therefore not be left in a vicious circle. It is necessary to act as quickly as possible to help poor decoders or children likely to become poor decoders. The questions are:

• which children need phonological training?
• should training address specific decoding deficiencies?
• are the effects generalized to understanding?

With regard to the training conditions (group size, type of training, duration, mode - individual or small group -) the tests and studies designed to determine the effectiveness of the various modalities have yielded very variable results.

The results of the main studies show that the effects of training in ‘phonological consciousness’ associating auditory and visual modalities has a marked statistical impact on development and a more moderate but statistically significant effect on reading and writing. Not only word identification but also understanding are enhanced. A more or less marked benefit with respect to reading was observed under all training conditions. The effects were positive for children learning to read normally, for children at risk of difficulties, and for those who had difficulties in nursery school or primary school class CP (6–7 years). Training also developed the writing skills of children in difficulty.

Thus, teaching phonological consciousness associated with the assembly route using intensive, specific and explicit training of small groups with similar needs promotes reading and writing in children experiencing decoding difficulties.

In France, preventive actions in the school environment were not inventoried in the context of this expert review and very few of those operations have given rise to studies published in the scientific literature.

A French training study using auditory and visual discrimination software was conducted on ‘poor decoder’ children in nursery school class GS (5–6 years). The training lasted 10 hours over 5 weeks at a frequency of 2 times 15 minutes daily, 4 days per week. It enabled the trained children to become better on average than the control children who decoded better at the outset. Another French controlled study addressed the effects of in-school training of 80 poor-decoder children in primary school class CE1 (7–8 years). Specific software was used for small groups. The study showed that trained children progressed more rapidly with respect to decoding than the untrained control group of poor decoders.

The software requires new rigorous evaluation in order to determine its specific effects and limits and the conditions for generalization of software use in the event that its beneficial effects are confirmed. The type of tool does not seem to replace a real teacher but to act as a complement.

However, the studies show that training in decoding does not improve or only improves a little the reading speed. The latter requires other types of training. According to the literature, the training recognized to be most effective with respect to fluidity (and hence speed of reading) is the technique of repeated reading: the letters, words and sentences read are repeated until a certain speed is achieved. Several authors have described the effects of such training on text reading speed (training of duration 6 minutes daily for 6 to 9 months).

In terms of prevention, it must be stressed that teaching adaptation is absolutely indispensable. The adaptations are designed to enable children to overcome their handicaps. The adaptations consist in taking into account the difficulties encountered by the child due to his/her disorder and enabling the child to learn in his/her preserved fields without being handicapped by his/her disorder (e.g. read the problems or give extra time for mathematics in the event of dyslexia, decrease the writing load in the event of dysgraphia, etc.). Descriptions of the adaptations are available in several forms (awareness-promotion CD-Rom, guides issued by the academies of Grenoble and Rennes for the use of teachers, dyslexic child follow-up book). In France, reflection on the nature of the adaptations and on the criteria required in order for children to benefit with a view to harmonizing the measures is indispensable.

Assessment of rehabilitation methods

There are a great variety of rehabilitation and training methods and some of them directly result from the various theories advanced to explain dyslexia. However, a large number of the recommended treatments for dyslexia have not undergone scientific study either from a theoretical viewpoint or from the viewpoint of evaluating the efficacy of treatment. The scientific criteria which enable treatment efficacy evaluation are not always totally fulfilled. The criteria include: studies of groups based on objective and quantified observations and supported by rigorous statistics (individual case reports do not constitute an objective evaluation); constitution of a control group which receives a placebo treatment of similar duration and intensity to that received by the test group; conventional procedure for a randomized, double-blind, controlled clinical trial; several independent studies conducted on a large number of subjects.

Most of the rehabilitation currently implemented is of the speech therapy type. Rehabilitation addresses: training the child with respect to phonological capability; reading rehabilitation with methods that are frequently different from those possible in class; setup of compensation strategies to enable the child to avoid the identified deficits.

While speech therapy rehabilitation as practiced in France is based on general principles derived from scientific knowledge acquired and validated in recent years, it has yet to undergo scientific assessment in the treatment of dyslexia. However, the fact that speech therapy is implemented individually with a specifically trained speech therapist enables development of an individualized rehabilitation program based on a precise assessment of the child’s skills and weaknesses. The program can be implemented under optimum conditions of interactivity enabling real-time adaptation to the child’s needs. That essential feature of speech therapist management can only, a priori, enhance efficacy compared to training programs administered to classes or groups as evaluated in the scientific literature. The international literature on training programs (computerized or not) applied to ‘poor reader’ children show that a degree of intensity is required in order to achieve a reasonable effectiveness (typically 4 to 5 sessions weekly). Given that intensity, good results can be obtained over relatively short rehabilitation durations (of the order of 6 weeks). Nonetheless, there is great variety in speech therapists’ practices. Those practices should therefore be evaluated and compared taking into account the modalities of their implementation (of which, intensity, duration and use or non-use of computer tools) in order to further identify good practices and in order to enable orientation of speech therapy practice. A computer program in French involving audiovisual training in identifying spoken and written syllables had a degree of efficacy. However, studies based on the criteria necessary for rigorous evaluation will be required in order to consider whether such training is of value as a complementary technique to speech therapy rehabilitation.

Among the various theories advanced to explain dyslexia, the theory in which the dyslexic (and dysphasic) subjects suffer from a deficit in temporal auditory processing has naturally given rise to training designed to rehabilitate auditory perception. This is the case of the US computer program Fast ForWord which includes training in discrimination of temporal auditory sequences and training in phonological skills. The special feature of the program is that it uses speech modified to render it more intelligible for children who have a temporal auditory processing deficit (short sounds and fast transitions amplified and prolonged in an adaptational manner). Nonetheless, the independent assessments of the program generated contradictory results and its efficacy on spoken and written language has yet to be clearly demonstrated.

The Tomatis method of auditory stimulation is based on a concept of audition that has no scientific basis. A meta-analysis of all the studies conducted did not show that the method had a positive effect.

The semiophonic method (or Lexiphone method) is an intensive rehabilitation method whose basic principle is auditory stimulation (‘parametric sound’) which is not based on any known scientific principle. The rehabilitation incorporates structured exposure to speech (syllables, words and pseudo-words), music and read texts, guided reading and writing. A clinical trial is underway in France.

Several rehabilitation methods for dyslexia address the visual capabilities and visual and attentional capabilities. Occlusion of one eye of children who were both dyslexic and had binocular instability was tested in the course of randomized, double-blind, controlled trials, which showed a significant improvement.

It has been suggested that wearing tinted glasses or lenses may improve the reading of dyslexic subjects with visual symptoms. A clinical trial showed modest beneficial effects on reading of the use of a transparency whose color was individually optimized for subjects suffering from visual stress.

A method known as ‘specific hemispherical stimulation’ consists in specifically stimulating the brain hemisphere which is presumed deficient in a given child by briefly (tachistoscopically) and repeatedly presenting words in the contralateral visual field. Experimental data suggest that the effects observed are non-specific (e.g. attentional).

The motor and proprioceptive systems have also been the subject of rehabilitation (motoricity, balance, archaic reflexes, proprioception). The data currently available are not sufficient to assess the efficacy of those methods in the treatment of reading disorders.

Several drugs and food supplements were not effective in clinical trials (methylphenidate, antihistamines, piracetam, polyunsaturated fatty acids).

No psychotherapy for dyslexia has been reported in the scientific literature. When a child presents with anxiety or depression or various behavioral disorders, in certain cases a difficulty in learning to read may result. But those disorders are of a very different nature to those encountered in the context of the definition of dyslexia. Nonetheless, may dyslexic children suffer from their failures at school and this may secondarily induce anxiety, depression or behavioral disorders. When that is the case, the disorders are to be diagnosed and appropriately managed in addition to the specific management of the learning disorder.

Individual management in clinical practice

Care consists in individual management of the child determined by precise evaluation of the deficits in written language, arithmetic and writing and the associated disorders (e.g. spoken language, affective or attentional disorders). In France, management is conducted by specialized professionals depending on the diagnosis (speech therapists, neuropsychologists, psychologists, psychomotricians, ergotherapists, etc.) and in the context of pluri-disciplinary devices and networks for the management of specific learning disabilities (reference centers, Rased, etc.). The objective of this expert review is not to analyze the practices in France, which is the responsibility of other organizations in the context of their missions, but to present a few principles for action based on analysis of the literature on the care field.

Irrespective of the care prescribed, it must also be associated with pedagogic management in order to enable the child to continue learning in his/her areas of skill, despite his/her disorder and in order to offer teaching adapted to the child’s requirements and possibilities in the deficient area.

Few scientific studies have reported objective data on the indications for speech therapy rehabilitation in the context of a written language disorder (At what age? For what disorder severity?), or on the orientations and instruments to be used or the frequency or duration.

The Anaes recommendations (1997) on speech therapy for specific written language disability essentially consist in recommendations based on professional agreements (with no scientific evidence). However, speech therapy rehabilitation is widely practiced and is of major value in enabling the child to alleviate his/her deficiency. The published data on the development of written language and the predictive factors, the increasing number of studies to evaluate specific training and the open studies on the benefits of intensive management of severely affected children will enable definition of at least some indications, orientations and practical conditions.

In the event of a specific disorder of written language acquisition, individual speech therapy rehabilitation is recommended as of primary school class CP (6–7 years) if a spoken language disorder persists or as of the end of CP if the initial adapted teaching response proves insufficient or in the event of signs of seriousness such as the absence of grapheme-phoneme or syllable matching (/b/a/ → /ba/), particularly if there is a familial history of language disorder or a personal history of delayed spoken language acquisition. Implementing reading and writing rehabilitation concomitantly seems preferable. The orientations are determined by the precise results of the individual evaluation of the deficient and preserved strategies, the cognitive functions underlying spoken language, phonological skills and visual processing, and quantitative and qualitative evaluation using standardized tests. Decoding and encoding by assembly and phonological skills constitute the first phase targeted in rehabilitation, if they are not sufficiently efficient and automatic. This is usual in most cases of dyslexia.

Visual, kinesthetic and semantic aids may be used to offset the auditory perceptual deficit. The orthographic stock is generally developed by differentiating the situations with no visual disorder and those with a visual processing deficit. In phonological dyslexia with no visual processing deficit, the decoding difficulties slow the constitution of the orthographic lexicon which is to be developed. In cases in which visual processing is disturbed, it is also rehabilitated to enable the child to enrich his/her orthographic lexicon. The use of that lexicon, once it has been developed, is to be promoted through writing then reading by working on the words in isolation or in context. The use of morphology is promising for dyslexic subjects with a view to helping the child develop orthographic knowledge independently of phonological skills. Lastly, fluid reading is developed by repeated reading and, above all, comprehension skills are worked on in order to bring the written comprehension level at least up to the spoken-language comprehension level. The project generally defines a limited number of axes for a given time period. The most urgent as a function of the child’s age and type of disorder take prevalence. For instance, the most urgent may be grapheme-phoneme matching and sound discrimination. The axes are reviewed by means of regular fractional quantitative and qualitative evaluations of the target deficient function.

Studies of the various types of training show that intensive training programs (half an hour daily, 4 days per week) over relatively short durations (5 to 10 weeks) procure specific benefits with respect to the trained function providing that they are precisely and specifically targeted on a deficient cognitive function. Those results may provide the grounds for reviewing current care practices. Intensive, specific, short-duration programs may be proposed for children recalcitrant to adapted pedagogic management and harmonized with the teaching which remains indispensable. Nonetheless, training on a precise function does not constitute the whole of speech therapy rehabilitation. The objective of the latter is to achieve functional reading and legible writing. The overall effects of rehabilitation need to be evaluated every 6 months to one year by using a new assessment employing similar tests to the initial assessment and enabling quantitative and qualitative assessment of the progress achieved and hence pursuit of the rehabilitation project. Depending on the time course of the child’s skills, the decision may be to discontinue rehabilitation (in the event of score normalization or score stabilization with functional reading and legible writing), or continuation of rehabilitation at a rate to be determined as a function of the objectives to be achieved (in the event of improvement without normalization or perfectly functional reading or legible writing), or pluri-disciplinary evaluation, for example by a reference center with a view to detecting associated disorders in the event of an inadequate course.

After the end of rehabilitation, adaptation in the school environment remains indispensable and is to be based on the handicap in terms of reading speed and dysorthography. A computer may be of value: word processing (in the event of associated writing disorders), spell check, vocal dictation (in the event of marked sequelae) in order to offer the child text reading by the computer and enhance the legibility of the child’s written production.

Writing disorders need to be precisely analyzed in order to ensure an appropriate response from the most pertinent professional (psychomotrician or ergotherapist). The aim is to determine, on the basis of the quantitative and qualitative results of specific tests, whether the disorder affects gesture coordination, perception and/or visual or visual and spatial production. Management may begin as of the end of nursery school, at the start of primary school if the disorders are severe and before age 8 years in the event of persistent disorders, before deviant strategies of letter sequencing become established. Like speech therapy rehabilitation, the management of handwriting is to be associated with appropriate teaching responses and is to be evaluated by progress assessment using standardized tests whose results will be compared with those of the initial assessment. Orthoptic management of a predominantly visual perception disorder may be indicated and the effects are to be evaluated. Learning word-processing or even vocal dictation depends on the comparative assessment determining the degree of persistent handicap given the school project.

The management of mathematical disorders has been very little studied. The instruments and trained professionals are only available in inadequate numbers. Nonetheless, disorders in the acquisition of the numerical code may be managed from the start of primary school, particularly since concomitant written language acquisition disorders are frequent, exacerbating the child’s failure.

The disorders associated with written language acquisition disorders and affecting the spoken language require rehabilitation since the links between the spoken and written language are very close. With a view to preventing dyslexia, any specific spoken language disorder must be taken into account. Rehabilitation has two objectives: improving speech and language but also preparing for written language (work on assembly and phonological consciousness).

The associated behavioral and affective disorders require an appropriate care project. The assessment of learning in any child consulting for attention deficit/hyperactivity disorder and the evaluation of behavior and attention in any child consulting for a written language acquisition disorder will enable management of the two disorders if they are associated. Anxiety and mood disorders are to receive psychotherapeutic care. Psychotherapies addressing the child’s psychological development are perfectly compatible with cognitive rehabilitation programs. The complementarity of the managements, providing that one does not exclude the other, addressing the whole child and his/her diversity at both cognitive and environmental relational level can in practice be implemented in the field with benefit for the child both with regard to the child’s adaptation to his/her difficulties and his/her compliance with rehabilitation.

Care coordination and family counseling are indispensable and are the responsibility of the child’s physician, trained referral professional and the various other professionals involved. Coordination is essential in order to define the objectives of the management programs, taking into account their feasibility and priority for the child, and also in order to assess the child’s progress and readjust the objectives in consequence.

The harmonization of the therapeutic project and teaching project constitutes another fundamental aspect of management which can be based on the specialized teachers in the help and support networks and National Education physicians and psychologists. The choice of scholastic adaptation as a function of the child’s course must enable the child to continue learning while minimizing the affective consequences of failure.

Overall, the variety of professional practices implemented in the context of care and the need for health-education coordination call for an analysis, under the responsibility of the High authority for health, with a view to compiling good practice recommendations.

Enhancing detection, screening and prevention

The knowledge acquired on learning disabilities should be made available to the public, particularly the parents and professionals concerned. The knowledge will contribute to facilitating detection of children experiencing learning difficulties.

In school, teachers are the first and best placed people to identify children presenting with learning difficulties. An understanding of the nature of specific learning disabilities and their presentations should be incorporated in initial teacher training as should new knowledge with respect to educational practices.

Enhanced management

Widening our knowledge of scholastic learning and learning disabilities

The research orientations proposed are designed to develop an enhanced understanding of scholastic learning (reading, writing, arithmetic, etc.) and learning disabilities, in particular enhanced understanding of the causes of dyslexia and estimation of the frequency of the various specific learning disabilities in France.

Introduction: the Inserm collective expert review approach

Jeanne ETIEMBLE explained that the collective expert review approach consists in reviewing the international literature on a given subject. The review is a response to a request formulated by a sponsor, in the present case, the RSI (social regimen for free-lance workers, formerly CANAM), in the context of a program focusing on child health. After the expert reviews on obesity and mental disorders, the third collective expert review, considered herein, addressed the specific learning disabilities, namely dyslexia, dysorthography and dyscalculia. The group of experts - whose composition was validated by the Directorate General of Inserm – consisted of 11 specialists. The group issued a report consisting in three sections: analysis of the international literature; synthesis of the experts’ work; and the experts’ recommendations. Jeanne ETIEMBLE stressed the distinction between the analysis of the international literature implemented by the experts and a status report on French practices in the field considered. The latter was not included in the expert review’s scope and is rather a matter for the High Authority of Health. The collective expert review on specific learning disabilities in children is now at the stage when it is rendered public and the present meeting is intended to enable all parties to express their feelings, in particular to prevent any misunderstandings.

The subject: ‘Dyslexia, dysorthography and dyscalculia’ was selected in order to define a field of expert review less broad than scholastic failure for which considerably greater resources would have been necessary. While dyslexia is a frequent concern, there are numerous erroneous ideas with regard to the subject. It would therefore seem of value to review the question. Dysorthography and dyscalculia have been the subject of markedly fewer articles. Jeanne ETIEMBLE agreed that dysphasia, a specific spoken language disorder, constitutes such an important subject that it could in itself have been the subject of an expert review. The same observation applies to dyspraxia. However, the experts decided to consider those two disorders in the context of ‘disorders associated’ with dyslexia, dysorthography and dyscalculia.

Claire MELJAC (psychologist, Hôpital Sainte-Anne) noted that dysorthography, which is associated with dyslexia in most cases, was addressed indirectly and was thus present in the expert review. However, no significant findings have emerged with respect to dyscalculia and certain results on the subject are far from certain. Claire MELJAC therefore wondered why dyscalculia was included in the title of the expert review.

Ariel CONTE (CORIDYS) considered that the framework of the expert review, restricted to dyslexia, dysorthography and dyscalculia, was liable to restrict its public health interest when the teaching hospitals (CHU) statistics showed that few cases of homogeneous dyslexia existed. Won’t decision-makers have a narrow vision of specific learning disabilities?

Principal findings of the collective expert review

Recommendations

Methodology

An Inserm collective expert review¹⁴ sheds scientific light on a given subject in the field of health on the basis of a critical analysis and synthesis of the international scientific literature. The expert reviews are implemented at the request of institutions wishing for access to recent research data pertinent to their decision-making process with respect to public policy. An Inserm collective expert review is to be considered as an initial stage that is necessary but most frequently not sufficient to result in decision-making. The conclusions of the collective expert review contribute to, but cannot replace, debate between the professionals involved or society debate if the questions addressed are particularly complex and sensitive.

At the request of a sponsor, the Inserm collective expert review may be accompanied by an ‘operational’ expert review addressing application of the knowledge and recommendations and taking into account contextual factors (existing programs, structures, players, training, etc.). The latter type of expert review elicits contributions from the players in the field able to respond to the feasibility aspects, representatives of the administrations or institutions responsible for promoting applications in the field involved, experts having contributed to the reviews, and representatives of patient associations. The sharing of varied cultures and experience enables a complementary approach to the collective expert review in an operational framework. Moreover, a variety of work (recommendations for good practices, public hearings, etc.) implemented under the auspices of the High Authority for Health (HAS) may follow an Inserm collective expert review.

Collective expert review has been an Inserm mission since 1994. Some sixty collective expert reviews have been implemented in numerous health fields. The Institute guarantees the conditions under which the expert review is implemented (exhaustiveness of the document sources, qualification and independence of the experts, transparency of the process).

The Inserm Center for Collective Expert Reviews organizes the various stages of collective expert review from the initial problem statement through to communication of the report, with the assistance of Inserm departments. The Center team, consisting of engineers, researchers and a secretariat, implements the document searches, logistics and chairing of the expert review meetings. The team contributes to the scientific writing and to compiling the expert review products. Regular exchanges with other public organizations (EPST) implementing the same type of collective expert review have enabled similar procedures to be set up.