Soutenance mémoire
La persistance du retard de langage précoce
Late-talkers’ language outcome by school entry : A retrodictive study of predictors
Developmental difficulties in late-talkers
One way to enhance the prediction of language outcomes in LTs could be to consider their overall early developmental profile. Numerous studies have shown that LTs tend to have co-occurring difficulties in other developmental domains such as motor development, sleep, inattention behaviors and psychosocial adjustment. However, the possible contribution of these developmental difficulties to LD persistence has been sparsely examined. In this paper, we examine the early developmental profile of LTs and how it may additionally inform the prognosis of LD persistence.
Motor development
Problems in fine and gross motor development (Bishop & Edmundson, 1987b; Haynes & Naidoo, 1991; Hill, 2001; Tallal, Dukette & Curtis, 1989) have been observed in children with early LD. They tend to have more difficulties with hand coordination in fine motor tasks such as finger tapping or peg-moving (Bishop & Edmundson, 1987b; Hill, 2001), with balance (Hill, 2001), and with early coordinated walking (Haynes & Naidoo, 1991). Indeed, less than 50% of children with early LD walked without help at 15 months, and many still did not walk on their own at 18 months, rates much higher than what is expected for chronological age (Haynes & Naidoo, 1991). More importantly, Bishop and Edmundson (1987a; 1987b) found that early motor difficulties differentiated children with persistent versus transient LD from ages 4 to 5.5 years.
Sleep
Early sleep problems have also been reported in LTs. For instance, 2-year olds with LD were reported to have more trouble sleeping through the night and to get out of bed more often than children with typically developing language (Carson, Klee, Perry, Muskina & Donaghy, 1998; Caulfield et al., 1989), but results are inconsistent (Tervo, 2007). Dionne and colleagues (2011) have further shown that early sleep consolidation at 6 and 18 months (i.e. the maturation of sleep patterns whereby children’s night-time sleep duration increases as daytime sleep duration decreases) differentiated LTs- from LTs+: LTs- had less mature sleep consolidation up to 18 months.
Inattention behaviors
Inattention behaviors have also been documented in young children with LD (Hart, 2004; Horwitz et al., 2003; Tervo, 2007). Horwitz and colleagues (2003) found more inattention problems reported by parents in children with LD between 18 and 39 months than in controls. Similarly, Tervo (2007) showed that 2-year-olds with both expressive and receptive LD had more inattention behaviors reported by parents than children with no delay. Inattention problems in LTs could stem from common neurological bases between inattention and LD (Webster & Shevell, 2004), or may cause problems in processing language input (Dionne & Ouellet, 2011). Alternately, LD may cause difficulties in regulating effortful attention (Kopp, 1992).
Psychosocial adjustment
LTs are more prone to internalizing problems, such as anxiety and depressive symptoms (Carson et al., 1998; Caulfield et al., 1989; Hawa & Spanoudis, 2014; Henrichs et al., 2013; Tallal et al., 1989; Tervo, 2007), externalizing problems, such as opposition and aggression (Caulfield et al., 1989; Hawa & Spanoudis, 2014; Henrichs et al., 2013; Horwitz et al., 2003; Qi & Kaiser, 2004) and social difficulties, such as social withdrawal and poorer social skills (Carson et al., 1998; Caulfield et al., 1989; Irwin, Carter, & Briggs-Gowan, 2002; Paul, 1993; Qi & Kaiser, 2004; Rescorla, Ross & McClure, 2007; Tallal et al., 1989; Tervo, 2007). Psychosocial difficulties appear more severe in LTs with co-occurring receptive LD (Carson et al., 1998; Rescorla et al., 2007; Whitehurst & Fischel, 1994). LD may hinder the regulation of emotions, behaviors and social encounters (Dionne, Tremblay, Boivin, Laplante & Pérusse, 2003; Kopp, 1989), or psychosocial difficulties could affect children’s receptiveness to language stimulation (Dionne & Ouellet, 2011).
In sum, many studies have shown that LTs have co-occurring difficulties in motor development, sleep, inattention behaviors and psychosocial adjustment, but it is unclear whether these difficulties uniquely predict the persistence of LD. Indeed, previous studies have only focused on few developmental indicators, and the vast majority of them consider LTs as a homogeneous group without taking LD persistence into account. This may mask meaningful early developmental differences in multiple domains between LTs+ and LTs- that could be used to inform their prognosis.
The retrodictive approach
Most reviewed studies considered LTs as a homogeneous group without taking LD persistence into account. There is a methodological alternative, the retrodictive approach (Bishop & Edmundson, 1987a; 1987b; Dionne et al., 2011; Woods, Rieger, Wocaldo & Gordon, 2014). This approach relies on a prospective longitudinal data collection of various aspects of development in a large sample. It uses the benefits of hindsight to predict differing outcomes in children with a similar initial condition from information collected prospectively. To study the predictors of the persistence of LD in LTs, a first step is to identify LTs at an early age. The second step is to follow them longitudinally to record their language outcome at a later age. The third step is to identify differing outcomes (e.g. transient versus persistent LD groups). The fourth step is then to go back to compare groups on developmental indicators present at onset. On that basis, one can study the early predictors of persistence or transience.
This approach has been rarely used despite its advantages in predicting differing outcomes from a similar initial condition. Bishop and Edmundson (1987a; 1987b) used this approach and were able to show that motor difficulties at 4 distinguished LTs with persistent versus transient delays between ages 4 to 5.5 years. Similarly, Dionne et al. (2011) demonstrated that poor sleep consolidation at 6 and 18 months predicted the persistence of LD from 18 months to age 5 years. Unfortunately, studies using this design also have only focused on single developmental indicators. A more inclusive approach assessing multiple developmental indicators may shed light on other predictors of LD persistence.
The first objective of this study was to document the early developmental profile of LTs regarding their expressive and receptive language, motor development, sleep consolidation, inattention behaviors and psychosocial adjustment. Specifically, the question addressed was:
1) What are the differences in early developmental profile between LTs and children with typical language development at 18 months of age?
The second objective was to investigate what predicts the persistence of language delay in LTs prior to school entry. The questions addressed were:
2) Using a retrodictive approach, what were the differences in early developmental profile between LTs- and LTs+ identified at age 5 years when they were 18 months of age?
3) What developmental indicators at 18 months uniquely predict LD persistence in LTs prior to school entry? Do these indicators add to the prediction of persistence over and above the initial language profile?
Method
Participants and procedure
Late-talkers (LTs) were drawn from the Quebec Newborn Twin Study (QNTS) (Boivin et al., 2012). The QNTS is a prospective longitudinal follow-up of a population-based birth cohort of 1324 twins (662 families). It conducted annual assessments of cognitive, behavioral and socio-emotional aspects of development starting at 6 months of age.
Twin cohorts are typically used to consider questions relevant to the genetic etiology of phenotypes. However, given the extent of the longitudinal documentation they typically collect, they are also used as convenience samples to address developmental issues (e.g. Dale, McMillan, Hayiou-Thomas & Plomin, 2014; Dale, Price, Bishop & Plomin, 2003). Moreover, twins appear to be particularly appropriate to study LD since the prevalence of LD is higher in twins and the language development process does not appear to be qualitatively different than the one of singletons (Mogford-Bevan, 1999; Rutter, Thorpe, Greenwood, Northstone & Golding, 2003; Thorpe, 2006).
All twins born between November 1995 and July 1998 in the greater Montreal area were recruited at delivery. Families who accepted to participate in the study were contacted by letter and by phone. To be included, infants had to be born without major medical conditions, to have available birth data, to speak French or English, and one of their parent also had to be fluent in either French or English. Informed parental consent was obtained before each data collection.
For the total QNTS sample, 50.8% of twins were boys; 37.6% were monozygotic twins; the mean birth weight was 2.5 kg and the mean 5-minute APGAR was 9. The mothers’ mean age at birth was 30.5. When the twins were 6 months old, 82% of mothers had a high school diploma and 16.3% of mothers did not have French or English as their main language (e.g. Arabic, German). When the twins were 18 months old, 91% of families were 2 parent families and the mean family income was between 40 000 and 50 000 CND. Family characteristics of the QNTS sample have been shown to be very similar to those of an epidemiological sample of singletons representative of the Quebec population during the same period (Montplaisir & Petit, 1997).
Identification of late-talkers and controls in the QNTS sample
The current study focuses on assessments at ages 18 months and 5 years. LTs were identified at 18 months within the QNTS sample based on an expressive vocabulary score ≤15th percentile (n=136) on an adapation of the MacArthur Communicative Development Inventories – Short Form (MCDI-SF; Fenson et al., 2000) (see description below). At age 5 years, they were divided into two subgroups based on expressive vocabulary scores on an adapted version of the Peabody Picture Vocabulary Test (PPVT; Dunn & Dunn, 1997; Dunn, Thériault-Wallen & Dunn, 1993) as used by Malenfant et al., (2012) (see description below). LTs with scores ≤25th percentile by age 5 years were deemed to have a persistent LD (LTs-; n=36), and LTs with scores >25th and ≤75th percentiles were classified as having had a transient LD (LTs+; n=44). LTs with scores >75th percentiles by age 5 years were excluded (n=12). Forty-four (44) LTs identified at 18 months were no longer part of the study at age 5 years.
Control groups of children with typical language development (TYPs) were also identified within the QNTS sample for comparison purposes. For group comparisons at 18 months, TYPs had to have scores ≥25th and ≤75th percentiles on the adapation of the MCDI-SF at 18 months (n=454). TYPs with scores <25th or >75th percentiles at 18 months were excluded to avoid diminishing or inflating differences across groups. For group comparisons at 5 years, TYPs identified at 18 months who had an expressive vocabulary score ≥25th and ≤75th percentiles on the adaptation of the PPVT at age 5 years were retained (n=180). TYPs with scores <25th or >75th percentiles at 5 years were excluded (n = 158). One-hundred-sixteen (116) TYPs identified at 18 months were no longer part of the study at age 5 years.
Measures
Expressive language
At age 18 months, expressive vocabulary was assessed with the an adaptation of the MacArthur Communicative Development Inventories – Short Form (MCDI-SF; Fenson et al., 1993). Parents indicated words the child could say (expressive vocabulary) in a 77-word checklist. Words said by the child were summed to represent expressive language. The MCDI-SF is widely used in LTs studies (e.g. Bishop, Price, Dale & Plomin, 2003; Dale et al., 2003; Horwitz et al., 2003; Irwin et al., 2002; Thal et al., 1991), and has a high internal consistency, excellent test-retest reliability and good concurrent validity (Fenson et al., 1993; Fenson et al., 2000). Scores were corrected for gestational age, the positively skewed distribution was corrected with a square-root transformation, outliers in the first percentile were winsorized (i.e. replacement of outliers with the next lowest or highest score), and scores were standardized to Z-scores (mean of 0 and standard deviation of 1) on the whole QNTS sample.
At age 5 years, expressive vocabulary was assessed with an adapted French or English version of the Peabody Picture Vocabulary Test (PPVT; Dunn & Dunn, 1997; Dunn et al., 1993) in which children were asked to name designated pictures. This version of the PPVT was previously shown to provide reliable scores (Malenfant et al, 2012). Correct items were summed and total scores were corrected for gestational age. The negatively skewed distribution was corrected with a log transformation, and scores were standardized to Z-scores on the QNTS sample.
Receptive language
Receptive vocabulary was assessed at 18 months with an adaptation of the MCDI-SF (Fenson et al., 2000) (see psychometric properties above). Parents indicated words the child could understand (receptive vocabulary). Words understood by the child were summed, and scores were corrected for gestational age and standardized to Z-scores on the QNTS sample.
Motor development
Parents were asked to answer 8 yes/no items developed by Poe (Baker, Keck, Mott, & Quinlan, 1993) to assess fine (e.g. has your baby ever made a line with a pen) and gross (e.g. has your baby ever walked upstairs by himself) motor development at 18 months-of-age. Attained motor milestones were summed to create a motor development score. The Poe scale has been widely used in large studies such as the National Longitudinal Survey of Youth in United States, the National Child Development Survey in England, and the National Longitudinal Survey of Children and Youth in Canada. Ordinal alpha for the scale was .87. Scores were corrected for gestational age, outliers in the first two percentiles were winsorized and scores were standardized to Z-scores on the QNTS sample.
Sleep consolidation
Mothers were questioned about their child’s consecutive night-time and daytime sleep duration at 18 months. The duration of consecutive night-time sleep was assessed in rounded hours ranging from < 4 hours, 4, 5, 6, 7, 8, 9, 10 to > 10 hours, and the duration of consecutive daytime sleep was assessed in rounded hours ranging from does not nap, ≤ 1 hour, 2, 3, 4 to > 4 hours. The “less than” and “more than” categories were rounded to the nearest whole value to compute a sleep ratio of day/night consecutive sleep duration reflecting child sleep consolidation. Measures of sleep consolidation based on ratio scores have been shown to reliably reflect developing sleep patterns (Bernier, Carlson, Bordeleau & Carrier, 2010; Dionne et al., 2011). Scores were corrected for gestational age, outliers in the first two percentiles were winsorized, scores were standardized to Z-scores on the QNTS sample, and scores were inversed so that higher scores represent a more mature sleep pattern.
Inattention behaviors
The Social Behavior Questionnaire (SBQ; Tremblay, Desmarais-Gervais, Gagnon, & Charlebois, 1987; Tremblay, Vitaro, Gagnon, Piché, & Royer, 1992) was used at 18 months of age to assess children’s inattention behaviors. Parents were asked to rate the occurrence of their child’s inattention behaviors in the last 12 months on 3 items (was easily distracted, had trouble sticking to an activity; was unable to concentrate, could not pay attention for long; was inattentive) on a three-point scale (0 = never, 1 = sometimes, and 2 = often). Scores derived from the SBQ have been shown to be reliable (Tremblay et al., 1987; Tremblay et al., 1992). Item scores were averaged to create a scale score. Cronbach alpha was .62. For the QNTS sample, outliers in the ninety-nine percentile were winsorized and scores were standardized to Z-scores.
Psychosocial adjustment
The SBQ was also used at 18 months to assess children’s behaviors on seven areas of psychosocial adjustment: prosocial skills (3 items), aggression (4 items), opposition (2 items), hyperactivity (5 items), general anxiety (3 items), separation anxiety (4 items), and depressive symptoms (4 items). Parents rated items on a three-point scale (0 = never, 1 = sometimes, and 2 = often). Item scores were averaged to create 7 scale scores. Cronbach alphas ranged from .51 (depressive symptoms) to .78 (prosocial skills). For the QNTS sample, outliers for aggression, general anxiety, and depressive symptoms were winsorized and scores were standardized to Z-scores.
Child and family characteristics
Child sex, zygosity, birth weight and 5-minute Apgar score at birth were derived from medical records. Data on education, age at birth, mother main language (French, English or other such as Arabic or German), family income (10-point scale, 1 being less than 5,000 CND and 10 being more than 80,000 CND) and family structure (1- or 2- parent family) were obtained via self-report in an interview with the mother at the 6- and 18-month assessments.
Table 2.1 shows descriptive statistics for child and family characteristics for LTs and TYPs at 18 months (top portion), and LTs-, LTs+ and TYPs at 5 years (bottom portion). Overall, groups were similar on child and family characteristics with the exceptions that a greater proportion of LTs were boys (60.3% VS 49.6%, p = .03) and came from a family with a lower income (average family income between 30K and 40K VS between 40K and 50K, p = .02) compared to TYPs. A smaller proportion of mothers of LTs- used French or English as their main language than both LTs+ and TYPs (72.2% in LTs- VS 97.7% in LTs+, p < .01; 72.2% in LTs- VS 92.2% in TYPs, p < .01).
Attrition rates
The overall attrition rate of the QNTS between 18 months and 5 years of age is 21.2%, which is typical of longitudinal studies of this scope. The attrition for the sub-sample of this specific study was 27.1% between our 2 time points, which is similar to the QNTS attrition rate (p = .31). Thirty-two percent (44/136) of LTs identified at 18 months of age were lost to the study by age 5. Twenty-five percent (116/454) of TYPs identified at 18 months were lost to the study by age 5 years. The participants lost to our study were more likely to have a mother who did not use French or English as her main language, and to come from a single parent family. They were also more likely to have weaker expressive and receptive vocabularies, slower motor development and sleep consolidation, and weaker prosocial skills when they were 18 months of age. Statistical strategies (see below) were therefore implemented to control for potential biases due to selective attrition.
Statistical analyses
Analyses were performed with Mplus version 7.3 (Muthèn & Muthèn, 2014). It enables the correction of standard errors estimates for non-normality and the non-independence of twin data when twins are considered as individuals with the Maximum Likelihood Ratio (MLR) estimator. It also enables the treatment of missing data with the Full Information Maximum Likelihood (FIML) by fitting the model to all non-missing data to avoid biases due to attrition (McCartney, Burchinal & Kristen, 2006). For all analyses, alpha threshold was set at .05.
Tests comparing means (linear regressions in Mplus) were performed to compare developmental profiles at 18 months between 1) LTs and TYPs at 18 months and 2) LTs-, LTs+, and TYPs identified later at age 5 years.
Binomial logistic regressions were performed in a given sequence to predict the persistence of LD in LTs at age 5 years. First, a model with developmental indicators as predictors was tested (model 1a). The developmental indicators entered in this model were selected based on p-value ≤.20 (Maldonado, & Greenland, 1993) in previous tests comparing means. Second, another model with initial language levels (i.e. expressive and receptive vocabularies at 18 months) as predictors of LD persistence was tested (model 1b). Third, a final model including both the language and developmental indicators at 18 months was tested to determine unique predictors of LD persistence (model 2). Models 1a and 1b were compared to model 2 with chi-square tests to determine the contribution of each set of predictors over and above the other.
Results
Descriptive statistics
Table 2.2 presents means for expressive and receptive vocabularies, motor development, sleep consolidation, inattention behaviors and psychosocial adjustment scores (prosocial skills, aggression, opposition, hyperactivity, anxiety, separation anxiety and depressive symptoms) at 18 months and for expressive vocabulary at 5 for groups identified at 18-months, and groups identified at age 5 years.
Differences in the early developmental profile of LTs and TYPs at 18 months
The left portion of Table 2.2 presents the results of tests comparing means on the developmental profile indicators at 18 months of LTs and TYPs. Regarding language, aside from having lower expressive vocabulary scores at 18 months, LTs also had significantly lower receptive vocabulary than TYPs, and the effect size was large (Cohen, 1988). On other development indicators, LTs had lower motor, sleep consolidation and prosocial skill scores at 18 months than TYPs, and the effect sizes were small to moderate.
Retrodictive differences in the early developmental profile of LTs-, LTs+ and TYPs at 18 months
The right portion of Table 2.2 presents the results of tests comparing means on the developmental profile indicators at 18 months of LTs-, LTs+, and TYPs identified later at age 5 years using the retrodictive approach. Regarding language, while LTs- and LTs+ did not differ on expressive vocabulary at onset, LTs- had lower receptive vocabulary than LTs+, and the effect size was large. On other developmental indicators, LTs- had lower motor scores and showed less opposition than LTs+. The difference between LTs- and LTs+ for sleep consolidation was marginally significant (p = .087) though the effect size was moderate. LTs- had lower sleep consolidation at 18 months.
Compared to TYPs, LTs- had lower motor skills and sleep consolidation. Prosocial skills and anxiety scores at 18 months were only marginally different between LTs- and TYPs though effect sizes were moderate. LTs-tended to have poorer prosocial skills (p = .064) and higher levels of anxiety (p = .059). Interestingly, LTs+ were also more oppositional than TYPs at 18 months.
Prediction of the persistence of language delay in LTs prior to school entry
Table 2.3 shows the results of binomial logistic regressions performed to predict LD persistence in LTs at age 5 years. Results of model 1a with developmental indicators as predictors (i.e. with p-value < .20 in previous tests comparing means: motor development, sleep consolidation, opposition, and aggression) explained 27% of LD persistence. Motor development and opposition were unique predictors in this model and aggression was marginally significant. Slower motor development, less opposition and marginally higher levels of aggression predicted a persistent LD in LTs. Results of model 1b with 18-month expressive and receptive vocabularies as predictors explained 29% of LD persistence at age 5. Receptive vocabulary was the only unique predictor of LD persistence.
|
Table des matières
Chapitre 1 : Introduction
1.1 Introduction générale
1.2 Le développement du langage
1.3 Le retard de langage précoce
1.3.1 L’identification du retard de langage précoce
1.3.2 L’étiologie du retard de langage précoce
1.4 La persistance du retard de langage précoce
1.4.1 L’identification de la persistance du retard de langage précoce
1.5 La prédiction de la persistance du retard de langage précoce
1.5.1 Pourquoi est-ce difficile de prédire la persistance du retard de langage précoce?
1.5.2 Pourquoi est-ce important de prédire la persistance du retard de langage précoce?
1.6 Les difficultés développementales concomitantes au retard de langage précoce
1.6.1 Difficultés motrices
1.6.2 Difficultés de sommeil
1.6.3 Difficultés psychosociales
1.6.3.1 Comportements associés au trouble déficitaire de l’attention/hyperactivité (TDAH)
1.7 L’approche rétrodictive
1.8 Les difficultés à l’âge scolaire des enfants avec un retard de langage précoce
1.8.1 Difficultés langagières
1.8.2 Difficultés académiques
1.8.2.1 Lecture et écriture
1.8.2.2 Mathématiques
1.8.3 Difficultés psychosociales
1.8.3.1 Difficultés sociales
1.8.3.2 Problèmes internalisés et externalisés et comportements associés au TDAH
1.9 Résumé et rappel des objectifs de la thèse
Chapitre 2: Late-talkers’ language outcome by school entry : A retrodictive study of predictors
2.1 Résumé
2.2 Abstract
2.3 Introduction
2.3.1 Developmental difficulties in late-talkers
2.3.1.1 Motor development
2.3.1.2 Sleep
2.3.1.3 Inattention behaviors
2.3.1.4 Psychosocial adjustment
2.3.2 The retrodictive approach
2.4 Method
2.4.1 Participants and procedure
2.4.1.2 Identification of late-talkers and controls in the QNTS sample
2.4.2 Measures
2.4.2.1 Expressive language
2.4.2.2 Receptive language
2.4.2.3 Motor development
2.4.2.4 Sleep consolidation
2.4.2.5 Inattention behaviors
2.4.2.6 Psychosocial adjustment
2.4.2.7 Child and family characteristics
2.4.3 Attrition rates
2.4.4 Statistical analyses
2.5 Results
2.5.1 Descriptive statistics
2.5.2 Differences in the early developmental profile of LTs and TYPs at 18 months
2.5.3 Prediction of the persistence of language delay in LTs prior to school entry
2.6 Discussion
2.6.1 The role of receptive language
2.6.2 The contribution of the developmental profile
2.6.2.1 Motor development
2.6.2.2 Sleep
2.6.2.3 Opposition
2.6.2.4 Psychosocial adjustment
2.6.3 Clinical implications
2.6.4 Strengths and limits of the study
2.6.5 Further directions
2.7 References
Chapitre 3: Persistent versus transient early language delay: language, academic and psychosocial outcomes in elementary school
3.1 Résumé
3.2 Abstract
3.3 Introduction
3.3.1 Stability of preschool language skills into elementary school
3.3.2 Academic achievement in elementary school in children with early LD
3.3.3 Psychosocial adjustement in elementary school in children with early LD
3.4 Method
3.4.1 Participants
3.4.1.1 The Quebec Newborn Twin Study
3.4.1.2 The current study
3.4.2 Identification of early LD and LD persistence
3.4.3 School-age outcome measures
3.4.3.1 School-age language
3.4.3.2 Academic achievement
3.4.3.3 Psychosocial adjustment
3.4.3.3.1 Behavior problems
3.4.3.3.2 Social problems
3.4.4 Statistical analyses
3.5 Results
3.5.1. School-age language
3.5.2 Academic achievement
3.5.3 Psychosocial adjustment
3.5.4 Posthoc analyses controlling for family characteristics
3.6 Discussion
3.6.1 Stability of language skills into the elementary school years
3.6.2 Differing trajectories of academic achievement
3.6.3 Psychosocial adjustment problems in elementary school regardless of persistence
3.6.4 Hypotheses for later difficulties
3.6.5 Clinical implications
3.6.6 Strengths and limits
3.7 References
Chapitre 4 : Conclusion
4.1 Rappel des objectifs et des résultats de la thèse
4.2 Discussion des résultats
4.2.1 Les difficultés à l’âge préscolaire
4.2.2 Les difficultés à l’âge scolaire
4.2.3 L’hypothèse du rétablissement illusoire
4.2.4 Retour sur les hypothèses étiologiques
4.2.5 La prédiction de la persistance du retard de langage précoce
4.3 Les implications pour la recherche
4.3.1 L’identification précoce du retard de langage à 18 mois
4.3.2 La prise en considération de la persistance et l’approche rétrodictive
4.3.3 La prise en considération de plusieurs aspects du développement
4.4 Les implications pour la clinique
4.4.1 L’identification précoce du retard de langage et la prédiction de sa persistance
4.4.2 La prise en considération de plusieurs aspects du développement
4.5 Les forces et les limites de la thèse
4.5.1 Le recours à un échantillon populationnel
4.5.2 L’utilisation d’un échantillon de jumeaux
4.5.3 L’évaluation des habiletés langagières
4.6 Pistes de recherches futures
4.7 Conclusion générale
Références
Télécharger le rapport complet
