School-age outcomes of late-talking toddlers: Long-term effects of an early lexical deficit

Abstract

Background
Early intervention for children identified as late talkers (LTs) at the age of 24 months is still a controversial issue in research and clinical routine. Previous studies have shown inconsistent results regarding predictors of early lexical deficits on school‐age outcomes of late‐talking toddlers.

Methods
In a five‐wave follow‐up study we investigated various aspects of language and literacy abilities in 39 German‐speaking third‐graders who had been identified as late talkers (LTs) at the age of 24 months, compared to 39 typically developing children (TDC) also attending the third grade. The duration of auditory sensory memory was examined at the age of four years using mismatch negativity (MMN) of tones – an event‐related potential not confounded by any language skill. In addition, the predictive value of memory performance was examined in a longitudinal perspective.

Results
Overall, LTs scored within normal range in language and literacy assessments. However, LTs differed from TDC in vocabulary size, verbalization of semantic relations, non‐word repetition and spelling. The findings can be explained by phonological working memory. The duration of auditory sensory memory and spatial working memory did not account for any variance.

Conclusions
LTs sustain persistent phonological processing limitations even though their native language and literacy performance lay within the normal range at school‐age. Further research on second language acquisition, academic achievements and the efficacy of early intervention in late‐talking toddlers is recommended.
This article is protected by copyright. All rights reserved.

Full text

Accepted Article
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/desc.12826
This article is protected by copyright. All rights reserved.
Article type : Paper
School-age outcomes of late-talking toddlers: Long-term effects of an early lexical
deficit
Nicola Grossheinrich1,2,3, Gerd Schulte-Körne1, Peter B Marschik4,5,6 Stefanie Kademann1,7,
Waldemar von Suchodoletz1, Steffi Sachse1,8
1 Department of Child and Adolescent Psychiatry and Psychotherapy, University Hospital of the LMU Munich, Germany
2 Department of Child and Adolescent Psychiatry, Medical Faculty and University Hospital Cologne, University of Cologne, Germany
3 Department of Social Sciences, Institute of Health Research and Social Psychiatry, Catholic University of Applied Sciences of North
Rhine-Westphalia, Cologne, Germany
4 iDN - interdisciplinary Developmental Neuroscience, Institute of Physiology, Medical University of Graz, Austria
5 Child and Adolescent Psychiatry and Psychotherapy, University Medical Center Goettingen, Goettingen, Germany
6 Center of Neurodevelopmental Disorders (KIND), Department of Women's and Children's Health, Karolinska Institutet, Stockholm,
Sweden
7 Little Scientist House, Berlin
8 Institute of Psychology, University of Education Heidelberg, Germany
Corresponding author:
Nicola Grossheinrich
Catholic University of Applied Sciences of North Rhine-Westphalia
Wörthstraße 10
50668 Köln
Germany
phone: + 49 221 7757 390
e-mail: n.grossheinrich@katho-nrw.de
e-mail: nicola.grossheinrich@uk-koeln.de
Running title: School-age outcomes of late talkers

Accepted Article
This article is protected by copyright. All rights reserved.
Conflict of interest statement: No conflicts declared.
Research Highlights
 Third-grade language and literacy performance of former late talkers (LTs) lay within
the normal range.
 The LT group differs from typically developing children in vocabulary, verbalization
of semantic relations, non-word repetition and spelling at third grade.
 Differences between both groups can be predicted by phonological working memory.
 Ultimately, LTs sustain persistent phonological processing limitations even though
their native language and literacy performance lay within the normal range at school-
age.
Abstract
Background: Early intervention for children identified as late talkers (LTs) at the age
of 24 months is still a controversial issue in research and clinical routine. Previous studies
have shown inconsistent results regarding predictors of early lexical deficits on school-age
outcomes of late-talking toddlers.
Methods: In a five-wave follow-up study we investigated various aspects of language
and literacy abilities in 39 German-speaking third-graders who had been identified as late
talkers (LTs) at the age of 24 months, compared to 39 typically developing children (TDC)
also attending the third grade. The duration of auditory sensory memory was examined at the
age of four years using mismatch negativity (MMN) of tones – an event-related potential not

Accepted Article
This article is protected by copyright. All rights reserved.
confounded by any language skill. In addition, the predictive value of memory performance
was examined in a longitudinal perspective.
Results: Overall, LTs scored within normal range in language and literacy
assessments. However, LTs differed from TDC in vocabulary size, verbalization of semantic
relations, non-word repetition and spelling. The findings can be explained by phonological
working memory. The duration of auditory sensory memory and spatial working memory did
not account for any variance.
Conclusions: LTs sustain persistent phonological processing limitations even though
their native language and literacy performance lay within the normal range at school-age.
Further research on second language acquisition, academic achievements and the efficacy of
early intervention in late-talking toddlers is recommended.
Keywords: late talker, language impairment, phonological working memory, vocabulary,
MMN
Introduction
At the age of two, the productive vocabulary is one measure of linguo-cognitive
capacities that can be used to reliably identify a delay in language development. This delay,
called late talking in children, is generally defined as having both an active vocabulary of
fewer than 50 words and a lack of word combinations at the age of 24 months. Whereas some
researchers have mooted this type of delay as a potential precursor of specific language
impairment (SLI, e.g. Moyle, Stokes, & Klee, 2013), others consider late talking simply as a
variant of normal vocabulary development (e.g. Szagun, Steinbrink, Franik, & Stumper,
2006). In order to come to a more precise understanding of this phenomenon, there is a need

Accepted Article
This article is protected by copyright. All rights reserved.
for additional comprehensive longitudinal studies evaluating the significance of language
delays in late-talking toddlers.
Language and literacy development
Numerous small-scale longitudinal studies have been carried out to investigate the
impact of early language delays on further language development in late talkers (LTs). Most
studies have focused exclusively on age groups up to preschool-age (e.g. Fernald &
Marchman, 2012; Girolametto, Wiigs, Smyth, Weitzman, & Pearce, 2001; Marschik,
Einspieler, Garzarolli & Prechtl, 2007; Moyle, Weismer, Evans, & Lindstrom, 2007;
Rescorla, Dahlsgaard, & Roberts, 2000; Rescorla & Turner, 2015; Sachse & von
Suchodoletz, 2008; Thal, Miller, Carlson, & Vega, 2005). Language outcomes at ages 4 and 5
indicate that – in small sample sizes, in which LTs are thoroughly diagnosed and well-
selected – a large proportion of LTs scored within the normal range (see Rescorla, 2011); a
result which was supported by epidemiological studies at age 4 (Dale, Price, Bishop, &
Plomin, 2003, Hammer et al., 2017; Reilly et al., 2010).
Nevertheless, although LTs score in the normal range later on, they generally show
lower language abilities than typically developing children (TDC) in various language
assessments (Moyle et al., 2007; Thal et al., 2005), even up to primary school age (Paul,
Murray, Clancy, & Andrews, 1997, small-scale).
To date, few studies have investigated LTs until school-age and the results of those
that have are inconsistent, such that the underlying mechanism of LTs’ lower performance
remains unclear. For example, Rice, Taylor and Zubrick (2008) examined language abilities
of 128 former LTs who scored within normal range on all language tests at age 7 but
exhibited lower scores on specific speech-language assessments such as receptive
vocabulary, articulation, and grammar. Using a latent variable approach, Bornstein, Hahn,
Putnick and Suwalsky (2014) found evidence for language skill stability in 324 children,

Accepted Article
This article is protected by copyright. All rights reserved.
which was stronger from age 4 to adolescence than from ages 20 months to 4 years. This
finding was confirmed in a large-scale study by Dale, McMillan, Hayiou-Thomas and Plomin
(2014) who found that recovered late-talking children matched by TDC did not differ from
the latter in language and reading development at ages 7 and 12. Nevertheless, the recovered
LTs group was below the mean for the total sample - as a consequence continuing monitoring
was recommended.
Other studies that report school-age outcomes including reading and spelling
assessments found relationships between early language assessments (as expressive
vocabulary size) and reading outcomes prior to school enrolment (Hammer et al., 2017, large-
scale), at early adolescence in TDC (Bartl-Pokorny et al., 2013) and in population-based
studies (Bleses, Makransky, Dale, Hojen, & Ari, 2016; Lee, 2011).
In late-talking toddlers, reading was not affected at age 9 (Rescorla, 2002) in a small-
scale study. Surprisingly, subsequent reading was found to be diminished at age 13 in the
same sample (Rescorla, 2005). In contrast, Paul and co-authors (1997) reported that reading
and spelling (among receptive language) in 32 LTs were comparable to TDC, while LTs
displayed poorer performance on expressive language scales at second grade.
In sum, language and literacy skills of former LTs lay within the normal range at early
school-age, although diminished language and literacy performance could still be observed in
LTs, in contrast to TDC. The type and extent of reduced language and literacy abilities vary
across studies, leading to an inconsistent pattern of results. These inconsistencies could derive
from variations in the inclusion criteria of LTs and/or different language/literacy assessments
for specific age groups within the framework of diverse language backgrounds and language
characteristics. Moreover, there is concern that small-scale studies are underpowered and
therefore lead to differing findings (as only differences of large effect sizes could reveal
significance). On the opposite, in population-based studies (mostly) a small selection of

Accepted Article
This article is protected by copyright. All rights reserved.
assessments is reported (usually) excluding receptive language skills and/or working memory
achievements (leading to different findings in population-based studies in contrast to small-
scale studies). In particular, working memory achievements seem to be strongly related to
language and literacy development.
Working memory and language/literacy development
A central issue of debate is the question of whether working memory aberrations in
language and literacy performance are domain-general or domain-specific (for an overview
see Henry & Botting, 2017). According to domain-specific difficulties in language and
literacy achievements (e.g. Estes, Evans & Else-Quest, 2007), the phonological processing
limitation hypothesis (e.g., Shankweiler & Crain, 1986) can be applied stating that reduced
phonological processing interferes with higher-order cognitive processing. The relationship is
often discussed using the phonological loop as the verbal information processing unit of the
Baddeley & Hitch working memory model (1974) which additionally encompasses the
visuospatial sketchpad (for visuospatial material) and a central executive component
(coordinating the phonological loop system and the visuospatial sketchpad system). Initial
stages of information processing are described in Cowan’s working memory model (1995)
who postulates a first stage of a “sensory” auditory store, where information is held passively
up to 400 ms before being transferred to a second, longer-lasting store for further processing.
Interestingly, storage capacity and verbal information encoding speed seem to be related to
the development of vocabulary and grammar (e.g. Adams & Gathercole, 2000; Botting &
Conti-Ramsden, 2001).
In contrast to the domain-specific hypothesis of working memory, the domain-general
hypothesis (e.g., Swanson & Ashbaker, 2000) proposes that children with literacy difficulties
demonstrate more general executive deficits (and not isolated phonological processing
limitations). This domain-general hypothesis presumes that under high-demand conditions,

Accepted Article
This article is protected by copyright. All rights reserved.
the executive system draws resources from the phonological system, the visual-spatial system
or the general system (i.e., the long-term memory, see Baddeley & Logie, 1999). In line with
the domain-general hypothesis, more general working memory deficits become evident in
language disordered children (Lukács, Ladányi, Fazekas & Kemény, 2016; Vugs, Cuperus,
Hendriks, & Verhoeven, 2013).
Interestingly, one German study addressed this issue and found evidence for the
domain-specific hypothesis in German-speaking children with spelling difficulties, while the
domain-general hypothesis was proven for German children with reading difficulties
(Brandenburg et al. 2015). The domain-specific result was explained by the German
orthography which is considered more transparent in contrast to the English orthography. In
detail, German orthography is considered transparent based on the grapheme-to-phoneme
correspondence (relevant in reading), whereas the phoneme-to-grapheme correspondence
(relevant in spelling) is considered less transparent, thereby implying an increased need for
phonological processing. Thus, typically developing children learning consistent
orthographies are in general ceiling in reading accuracy at a time when children of more
oblique orthographies (especially English) are still struggling (Landerl et al., 2013).
Performance of phonological working memory is generally assessed by the repetition
of a series of tones, syllables, words, or numbers of increasing length. Specifically, reduced
non-word repetition (NWR) performance has been considered as a marker of SLI and is
assumed to be predictive for language development (Botting & Conti-Ramsden, 2001; Conti-
Ramsden, Botting, & Faragher, 2001; Gathercole & Baddeley, 1990). More importantly,
Stokes and Klee (2009) found deficient NWR performance to be related to late-talking
toddlers aged 24 to 30 months, suggesting that phonological processing is limited in LTs.

Accepted Article
This article is protected by copyright. All rights reserved.
Briefly, NWR is a powerful tool for assessing working memory. It can be used to
identify children with language impairments, not dismissing the notion that repetition
accuracy of non-words also depends on lexical and sub-lexical properties (Coady & Evans,
2008). This means NWR is confounded by linguistic abilities and is a pure assessment not of
auditory memory but of phonological working memory instead.
In contrast to NWR, mismatch negativity (MMN) of tones is an adequate assessment
for pure auditory memory (Näätänen, Jacobsen, & Winkler, 2005) widely used in language
and literacy research to determine auditory discrimination accuracy and auditory sensory
memory (Bishop, 2007).
Aims of the present study
The aim of the present study was to investigate long-term school-age outcomes of late
talking toddlers raised in a transparent language environment (first research question) and to
evaluate the predictive power of different memory domains (auditory sensory memory,
phonological working memory and spatial working memory) for language and literacy
outcomes (second research question). Phonological working memory and spatial working
memory were assessed which both require central-executive capacities. If both domain-
specific scores would be reduced we assumed a domain-general limitation of the working
memory while a domain-specific reduction is in favour for the domain-specific hypothesis.
For novelty reasons, language and literacy performance of well-selected LTs (in contrast to
TDC) were examined in a five-wave longitudinal study at third grade.
According to the first research question, we specifically assume differences in
vocabulary and the verbalization of semantic relations in LTs in contrast to TDC at third
grade as late talking is defined as a lack of active vocabulary which might be an underlying
mechanism later on. To date, the verbalization of semantic relations has not been investigated
in LTs at school-age. However, it is obvious that semantic relations depend on the amount

Accepted Article
This article is protected by copyright. All rights reserved.
and pattern of lexical encodings. Indeed, semantic networks are proven to be different
between LTs and TDC, with more semantically related words and stronger semantic
associations between words in TDC (Beckage, Smith, & Hills, 2011). Additionally, in line
with the results of Brandenburg and co-authors (2015), we hypothesised spelling aberrations
but not reading difficulties in former German-speaking LTs compared with TDC.
According to the second research question two different data analyses were conducted
to examine whether the lexical deficit is associated with reduced working memory capacity.
First, the duration of auditory sensory memory (MMN of tones) between LTs and TDC was
studied cross-sectionally in relation to the language and literacy performance of third-grade
children. Second, the influence of different working memory domains (auditory sensory
memory, phonological working memory and spatial working memory) was examined
longitudinally in order to predict differences in achievements between LTs and TDC at
elementary school age.
As late talking is associated with a reduced auditory sensory memory duration
(Grossheinrich, Kademann, Bruder, Bartling, & von Suchodoletz, 2010; implying that word
length is a crucial aspect of the early lexical deficit), and as late talking is related to a reduced
phonological working memory capacity (assessed by NWR, Stokes & Klee, 2009), we
assumed that working memory aberrations in LTs are specific for the auditory domain.
Methods
In a five-wave follow-up, various aspects of language and literacy abilities in 78
German-speaking children (32 females) were assessed. Initially, children with and without a
language delay took part in a longitudinal study beginning at 2 years of age (2;1 years, T1),
with follow-ups at the ages 3 (3;1 years, T2), 4 and a half (4;7 years, T3), shortly before
school enrolment (5;10 years, T4) and in the third grade of elementary school (after 26

Accepted Article
This article is protected by copyright. All rights reserved.
months of education, T5). For ethical reasons, all children could have any speech-/language
intervention starting at age three (as it is usual in any developmental study). Interventions
differed concerning methods, length and the child’s age at commencement.
Participants
Thirty-nine LTs (age 9;4 + 0;4, 15 females, 65% of the initial sample) and 39 TDC
(age 9;1 + 0;1, 17 females, 83% of the initial sample) were reassessed in the second half of
third grade (whereof two participants attended the first half of fourth grade). There was a total
attrition rate of 25%, and the LTs who dropped out did not differ significantly from the
remainder for all language tests at T1. All children displayed a normal non-verbal intelligence
score (IQ > 85) composed of the Snijders-Oomen Non-verbal Intelligence Test (SON-R 2 ½-
7; Tellegen, Winkel, Wijnberg-Williams & Laros, 1998, reliability: 87.-91.) which was
assessed at ages 3 and 5 (mean value). No group differences were observed for gender
(Fisher’s exact test, p = .82), handedness (Fisher’s exact test, p = .15), non-verbal
intelligence (T (76) = 1.44, p = 0.15), maternal education (U = -.85, p = .39) or months of
school education (T (76) = -1.13, p = .26). The groups (LTs vs. TDC) differed in age (i.e. LTs
were older than TDC; T (76) = -3.57, p < .01, Table 1).
The sample was recruited via birth announcements in a local newspaper, which at the
time of recruitment automatically published all birthdates. Families with 23-month-old
children thus received a letter asking them to participate in the study.
Children were classified as LTs at 2 years of age based on two different sources of
information: (i) a German version of the parent questionnaire (MacArthur-Bates
Communicative Development Inventories, CDI, Toddler Form; Fenson, Dale, & Reznick,
1993), the ELFRA-2 (Elternfragebogen fuer die Frueherkennung von Risikokindern,
ELFRA-2; Grimm & Doil, 2000, reliability: α = 84.-98.), and (ii) a standardized language test

Accepted Article
This article is protected by copyright. All rights reserved.
for 2-year-olds (Sprachentwicklungstest fuer 2-jaehrige Kinder, SETK-2; Grimm, Aktaş, &
Frevert, 2000, reliability: r = .62-.95). The language test is composed of two receptive and
two productive language subtests. Children with poor results in the parent questionnaire
ELFRA–2 (productive vocabulary < 50 words or vocabulary between 50 and 79 words and
morpho-syntactic deficits) and poor results in one of the standardized language subtests (z-
score < -1.5 at least in one subtest of SETK-2) were classified as LTs (n = 60). Children with
normal results in ELFRA-2 (vocabulary > 80 words and normal morpho-syntactic abilities)
and SETK-2 (z-score > -1 in all subtests) were defined as TDC (n = 47). Information about
developmental milestones, medical history (complications during pregnancy or birth, preterm
birth, history of otitis media or other ear disorders), and socioeconomic characteristics was
acquired by having the parents complete a questionnaire and take part in an anamnestic
interview. There were no complications during birth and pregnancy, no other critical
incidents and no chronic diseases reported for the participating children. All children had
normal hearing (measured by otoacoustic emission screening or audiometry) and non-verbal
abilities within the normal range (Münchner Funktionelle Entwicklungsdiagnostik, MFED,
subtest Perception; Hellbrügge, Lajosi, Menara, Schamberger & Rautenstrauch, 1985).
Children agreed to take part in the study and parents provided written informed
consent. The study was approved by the local ethics committee.
Assessments
Predictors (assessments of language and memory abilities) within the longitudinal
design are listed in Table S1. All test administrators were carefully trained psychologists and
blind to the initial group status (LTs or TDC) for all language, literacy and memory
assessments starting from T2.

Accepted Article
This article is protected by copyright. All rights reserved.
Language and literacy assessments
Predictors
Age-specific expressive and receptive standardized tests were administered at all
measurement points (SETK-2, Sprachentwicklungstest fuer drei-fuenfjaehrige Kinder, SETK
3-5, Grimm, Aktas & Frevert, 2001, reliability: r = 62.-89.). Assessments of language
abilities within the longitudinal design are listed in Table S1. Additionally, a test for
metalinguistic skills (phonological awareness) was applied at age 5 (Table S1).
Outcome
Third-grade language abilities were examined using a standardized battery that tests
language abilities in 5 to 10 year-old German-speaking children (Sprachstandserhebungstest
für Kinder im Alter zwischen 5 und 10 Jahren, SET 5-10, Petermann, Metz & Fröhlich, 2010,
reliability: α = .71-.91). The battery consists of 10 subtests depicting productive language
skills including lexical/semantic tasks (subtest active vocabulary, semantic relations),
grammar (subtest correcting incorrect sentences, subtest inflections - plurals of real and
pseudowords, subtest sentence combining), narrative abilities (subtest picture storytelling),
language comprehension (subtest text comprehension, subtest syntactic understanding) and
processing speed.
In the grammar subtests, children had to correct syntactic as well as morphological
mistakes within sentences (subtest correcting incorrect sentences), create plural forms of
existing words and pseudo-words (subtest inflections - plural forms), and generate sentences
based on two or more words (sentence combining). In order to measure language
comprehension, children were told a story and were asked questions about it afterwards
(subtest text comprehension). In addition, the test administrator read single sentences aloud
and the children had to re-enact the content with the help of toys (subtest syntactic

Accepted Article
This article is protected by copyright. All rights reserved.
understanding). Narrative abilities were evaluated via picture storytelling (subtest picture
storytelling).
Third-grade literacy skills were examined by testing reading speed (Salzburger
Lesetest, SLRT II, Moll & Landerl, 2010; reliability: r = .71-.86), reading comprehension
(Ein Leseverständnistest für Erst- bis Sechstklässler, ELFE 1-6, Lenhard & Schneider, 2006,
reliability: α = .92-.97) and spelling (Weingartner Grundwortschatz Rechtschreibtest für
dritte und vierte Klassen, WRT 3+, Birkel, 2007, reliability: α > .84).
Working memory assessments
Predictors
Assessments of memory abilities within the longitudinal design are listed in Table S1.
The assessment of phonological working memory comprised different tasks: non-word
repetition/NWR (SETK 3-5; T4, 5;10 years, reliability: r = .62-.89), digit span (Kaufman
Assessment Battery for Children, K-ABC, Melchers & Preuß, 2009, T4, 5;10 years,
reliability: .86), and a subtest which required children to reproduce words in the correct order
(word order, K-ABC, Melchers & Preuß, 2009; T4, 5;10 years, reliability: r = .84). In order
to evaluate whether memory skills other than phonological working memory are relevant for
the prediction of further language development, memory was also assessed for spatial
information (spatial working memory, K-ABC, Melchers & Preuß, 2009, T4, 5;10, reliability:
r = .74).

Accepted Article
This article is protected by copyright. All rights reserved.
Outcome
Non-word repetition (SET 5-10) was assessed and raw scores were reported for cross-
sectional analysis (Table 2). For regression analysis NWR was excluded from the principal
component analysis (PCA) and the resulting factor score.
Neurophysiological assessment
MMN is a well-established method for examining the second phase of auditory
sensory memory storage (Näätänen et al., 2005), which decays after 10 to 20 seconds and
thereby resembles auditory sensory memory when presenting non-verbal material. In order to
estimate the duration of auditory sensory memory, interstimulus intervals (ISIs) of different
lengths are used, especially for individuals suffering from disorders associated with memory
deficits and language impairments (for review see Bartha-Doering, Deuster, Giordano, am
Zehnhoff-Dinnesen, & Dobel, 2015).
In this study, the duration of auditory sensory memory was assessed at 4;7 years of
age (T3) using an oddball paradigm with varying ISI conditions presenting easily
distinguishable non-linguistic standard (1000 Hz) and deviant tones (1200 Hz). This approach
was reported first by Grau, Escera, Yago, and Polo (1998) and is described in detail in
Grossheinrich et al. (2010, Figure 1) as a time-saving oddball paradigm suitable for young
children.
In short, a control (ISI: 500 ms) and an experimental condition (ISI: 2000 ms) were
presented in separate blocks in a balanced order. During the EEG recording, children were
seated in an upright child-sized seat and were shown a video without sound. The children’s
guardian remained in the testing room, silently completing questionnaires. The total duration
of the experiment was approximately 42 min.

Accepted Article
This article is protected by copyright. All rights reserved.
The EEG was recorded using 20 Ag/AgCl sintered electrodes, and frontal mean
amplitudes of the MMN were analysed. The time window was based on running t-tests
against zero (p <.05 at 4 consecutive data points) from the evoked responses of the combined
group (LTs and TDC) for each ISI condition separately. For more details on the
approach/procedure see Grossheinrich et al. (2010).
Statistics
Multivariate analysis of variance (MANOVA) was performed for cross-sectional
analysis using a between-group design (LTs and TDC at third grade). Two subtests of the
SET 5-10 (storytelling and sentence building) were excluded from the analysis due to ceiling
effects (skewness < -1).
To account for confounded language and literacy assessments, a principal component
analysis (PCA) was applied to shed light on inconsistent patterns of results reported in
previous studies. PCA (direct oblimin rotation, delta = 0, creating almost independent
components with minimized intercorrelations) was conducted as appropriate multivariate
technique in order to extract important information out of various inter-correlated quantitative
dependent variables as a set of principal (nearly orthogonal) components.
For neurophysiological data, a three-factorial ANOVA was conducted with ‘group’
(LTs, TDC) and ‘performance category’ at third grade (factor score: positive values vs.
negative values) as between-subjects factors, and ISI (500 ms vs. 2000 ms) as within-subjects
factor. For the ‘performance category’, a split-half method was applied (as only a few third-
graders exhibited underachievement). The principal component comprising the differences
between LTs and TDC was chosen for the split-half method, given that LTs and TDC differ
especially in these assessments. Within this principal component, children exhibiting a

Accepted Article
This article is protected by copyright. All rights reserved.
negative (below-average value) or a positive (above-average) factor score were distinguished
into two groups.
For longitudinal analysis, a hierarchical stepwise regression was carried out in which
language and memory abilities were included to predict the factor score of the principal
component reflecting the variations between LTs and TDC. The factor scores were
recalculated again – this time without the phonological working memory value (NWR). To
account for missing data (5.5 % of all predictors), we used multiple imputation (MI)
technique and created 20 data sets. Estimated associations in each of the imputed datasets
differed because of the variation introduced in the imputation of the missing values. For this
reason, averaged estimated associations are reported.
In order to investigate the impact of memory skills on the differential outcomes at
third grade, memory scores were entered into the model first, followed by phonological
awareness and language performance. For regression analysis, the variable that had the
highest significant correlation with the factor score reflecting LTs and TDC differences was
entered first. Residuals were computed and the next most significant predictor was included
until no additional variance was explained.
In a second step, we controlled for the variables sex, IQ and intervention within the
analysis including each variable separately in the analysis first. Intervention was documented
dichotomously (children who received intervention: 31% vs. those who did not: 69%).

Accepted Article
This article is protected by copyright. All rights reserved.
Results
Behavioural results
Using Pillai’s trace, LTs demonstrated lower overall literacy and language scores
than TDC at third grade (V = .27, F (11,66) = 2.19, p = .03, partial η2 = .27). The between-
subjects effect was based on differences in vocabulary (F (1,76) = 5.7, p = .02, partial η2 =
0.7), verbalization of semantic relations (F (1,76) = 10.41, p < .01, partial η2 = .12),
phonological working memory (F (1,76) = 10.78, p < .01, partial η2 = .12) and spelling (F
(1,76) = 6.22, p = .02, partial η2 = .08) (Table 2). Although differences between LTs and
TDC were observed, the LTs’ performance was situated within the normal range (z > -1,
Table 2).
PCA (direct oblimin rotation, delta = 0) was conducted and yielded four components
with an eigenvalue > 1 (Table S2). The first component explained 26% of the variance and
was interpreted as ‘literacy-speed’ factor. The second component accounted for 14% of the
variance and reflected the variations between late-talking and control children (‘LTs vs.
TDC’). The third and fourth components each explained an additional 11% of the variance;
whereas the former was interpreted as ‘grammar’, the latter did not fit any explanation. The
structure matrix and the inter-correlation matrix are listed in the supplemental material (Table
S2 and Table S3).
Neurophysiological results
An ANOVA for mean MMN amplitude was performed with ‘group’ (LTs, TDC) and
‘performance category’ (third grade; above average: positive values; below average: negative
values) as between-subjects factors and ISI (500 ms vs. 2000 ms) as within-subjects factor.
The performance category was defined as the factor score of the second component (‘LTs vs.

Accepted Article
This article is protected by copyright. All rights reserved.
TDC’) comprising spelling, vocabulary, verbalization of semantic relations and phonological
working memory differences (see Table S2).
An interaction between ‘group’ and ‘ISI’ was found (F (1,58) = 7.04, p = .01). Post-
hoc ANOVAs yielded a significant group effect in the 2000 ms experimental condition
(F (1,58) = 5.7, p = .03) with an absent MMN in LTs (Figure 2). No group difference was
observed in the 500 ms condition (F (1,58) = 1.47, p = .23). No main effect was detected for
‘performance category’ (F (1,58) = 0.81, p = .78) or for the interaction with ‘group’
(performance category x group: F (1,58) = 0.34, p = .86) (Figure 2).
Longitudinal analysis
A hierarchical stepwise regression analysis was conducted including behavioural and
MMN data to predict third-grade group differences between LTs and TDC (‘LTs vs. TDC’
factor). In order to investigate the influence of memory on the second component (‘LTs vs.
TDC’), comprising vocabulary, verbalization of semantic relations, spelling and phonological
working memory, PCA was carried out again - this time excluding the phonological working
memory score (NWR). Memory and language achievements were entered into the analysis
beginning with the memory predictor that explains most of the variance. Residuals were
computed and the next significant memory predictor was included until no significant
additional variance explanation was found.
Two predictors were included in the model and accounted for 30 % of the variance.
Both predictors equate phonological working memory at the age of 5 years (NWR, 24 %, β =
.49; T = 4.66, p < .01 and word order, 6%, β = .24, T = 2.03, p = .04). Phonological
awareness was nearly significant and would account for additional 5% of variance (β = .22, T
= 1.73, p = .08; Table 3). Neither sex, nor intervention (dichotomously documented) nor IQ
altered the main results. Pearson correlations yielded a relationship between the ‘LTs vs

Accepted Article
This article is protected by copyright. All rights reserved.
TDC’ factor score of the third-grade performance and phonological working memory at 5
years (digit span, r (72) = .28, p < .05) but not for spatial memory (p > .3) or auditory short-
term memory (MMN 500 ms: p > .3; MMN 2000 ms: p > .2).
Power analysis
A post-hoc power calculation using G*power (G*Power 3.1; Faul, Erdfelder, Lang,
and Buchner, 2007) indicated power of approximately 95% for MANOVA and multiple
regression analysis, thereby indicating that our findings were identified correctly (β ≤ 5).
Discussion
The significance of late talking in toddlers is still the subject of discussion in research
and early intervention. In the present study, significant differences in language and literacy
skills of former LTs compared to those of TDC were found to persist in third-grade children.
Our findings will be discussed as related to phonological processing and can be explained by
phonological working memory performance (approximately 30%) while other memory scores
(e.g. auditory sensory memory, spatial working memory) did not account for any variance.
Nevertheless, the reported mean performance of the LT group lay within normal
range, which is in line with previous findings (for review see Rescorla, 2011). For ethical
reasons, children in this study were offered to receive intervention after their third birthday.
Such interventions were documented dichotomously and included in the regression analysis.
It can be assumed that specific speech and language therapy has at least a slight positive
effect on children’s further development (Law, Garrett & Nye, 2004). This effect applies to
all longitudinal studies addressing LTs and is one possible explanation for the relatively small
group differences later on.

Accepted Article
This article is protected by copyright. All rights reserved.
Long-term follow-up differences between LTs and TDC at elementary school-age
In the present study, we found that LTs differed from TDC in vocabulary,
verbalization of semantic relations, non-word repetition, and spelling.
Not surprisingly, the difference in vocabulary can be explained by the inclusion
criteria of less than 50 words for LTs and more than 80 words for TDC as active vocabulary
at the age of 2 years. However, it is remarkable that late-talking children still differ at third
grade which matches findings from school-age outcome studies (for review see Rescorla,
2011). In addition, in a large-scale study of 300 British children, infant vocabulary was found
to account for 16% of variance in later vocabulary (measured 5 years later, on average) which
is in line with our finding (Duff, Reen, Plunkett, & Nation, 2015). In conclusion, although
achievement of LTs at third grade falls within the normal range, alterations in vocabulary
acquisition are not limited to the third year of life but continue into later childhood.
The verbalization of semantic relations has not been investigated in LTs at school-age
thus far, although it is indisputable that semantic relations depend on lexical encodings. In
this study we found differences in verbalization of semantic relations in LTs compared to
TDC which might be related to an impairment of the word-object mapping needed to create
lexical encodings and which is reduced in late-talking toddlers (Weismer, Venker, Evans, &
Moyle, 2013). As NWR proves to be predictive for late talking (Stokes & Klee, 2009), the
altered performance in NWR identified in this study is in line with a permanent lexical
acquisition deficit which might be related to phonological processing limitations.
The aberrant performance in spelling is consistent with the results of Rescorla (2002)
who investigated LTs reading and spelling in primary school. Although Rescorla (2002)
subsumes spelling performance under reading abilities, the dictation score is noteworthy.
Hence, it seems that spelling is more affected than reading in late-talking children. One
possible explanation is that spelling (derived from dictation) requires more reliable access to

Accepted Article
This article is protected by copyright. All rights reserved.
the lexicon than does reading. Interestingly, when English native-speaking children with
language impairment were compared to non-impaired control children, reading accuracy and
spelling were diminished (accompanied by NWR deficits) while reading comprehension was
not affected (Baird, Slonims, Simonoff, & Dworzynski, 2011). In addition, one recent study
investigated children with isolated spelling disorder and found neurophysiological
abberations which could be explained by missing lexical encodings (Bakos, Landerl,
Bartling, Schulte-Körne & Moll, 2018). Thus, in transparent languages with regular
orthography – such as German – spelling and not reading accuracy might be the capability
more sensitive to NWR.
As a consequence of reduced phonological processing, LTs might be less able to
enlarge their vocabulary, which in turn results in altered semantic networks with impeded
access. Their grammar acquisition and literacy skills (in particular spelling) could be affected
by reduced phonological processing during schooling. Therefore, the aim of our second
analysis was to examine different memory scores and their contribution to the third-grade
performance differences between LTs and TDC.
The influence of memory on performance differences between LTs and TDC
The second research question investigated whether working memory aberrations in
language and literacy performance are domain-general or domain-specific. In particular, we
investigated whether phonological working memory or working memory in general has an
influence on third-grade language and literacy performance. We assessed the MMN which is
not confounded by any language skills, to estimate the duration of auditory sensory memory;
the latter was analysed since it was found that word length is predictive for the age at first
word production (Maekawa & Storkel, 2006). While differences in the MMN between LTs
and TDC were found, the MMN did not affect performance at elementary school-age.
Instead, memory assessments reflecting phonological working memory (‘word order’ of K-

Accepted Article
This article is protected by copyright. All rights reserved.
ABC and NWR) accounted for approximately 30% of the total variance in the final
regression model. One nearly significant result was explained by phonological awareness
(5%). It might be, therefore, that the early phonological working memory deficit in LTs turns
into reduced phonological awareness at pre-school age.
A question whose answer may help explain delayed vocabulary acquisition is whether
working memory deficits are limited to phonological processing or whether they are
characterized by more general memory impairments instead (Nickisch & von Kries, 2009).
Different durations of auditory sensory memory were reported between LTs and TDC at age
4 but not between late bloomers (children who overcame their early language delay before
age three) and children with continuing language difficulties (Grossheinrich et al., 2010). In
this study, no differences were observed between low and high performers at third-grade (see
Figure 2). The results suggest that children require just a minimal duration of auditory
sensory memory in order to start talking but the length of this duration is not predictive for
their performance in elementary school. Instead, our findings demonstrate that more specific
phonological working memory performance is needed for later language development.
Specifically, phonological working memory accounts for approximately 30% of the
variance explaining the differences between LTs and TDC, while auditory sensory memory
(assessed by the MMN) and spatial working memory were not associated with third-grade
literacy and language outcome.
A verbal storage deficit has already been reported for parents of late-talking children
in studies investigating NWR performance (Bishop et al., 2012). Consistent with altered
NWR performance in parents of LTs, deficits in NWR have also been demonstrated for
children with language impairments. Studies with children who were identified as LTs or
children recovered from a history of SLI reported significantly lower scores on NWR tasks
(Conti-Ramsden et al., 2001; Weismer et al., 2000; Thal et al., 2005). Phonological working

Accepted Article
This article is protected by copyright. All rights reserved.
memory deficits have also been found in the digit spans of children with SLI (Briscoe &
Rankin, 2009; Hick, Botting, & Conti-Ramsden, 2005; Hoffman & Gillam, 2004; Nickisch &
von Kries, 2009). More importantly, research has indicated that LTs perform poorly on NWR
tasks, and such poor performance is known to be strongly related to expressive vocabulary
size in two-year-olds (Stokes & Klee, 2009).
No relationship between third-grade outcome differences between LTs and TDC and
spatial working memory was observed. Some studies have reported comparable spatial
working memory skills between children with SLI and TDC (Archibald & Gathercole, 2006,
2007; Lum, Conti-Ramsden, Page, & Ullman, 2012) whereas others have observed
differences (e.g. Bavin, Wilson, Maruff, & Sleeman, 2005; Hick et al., 2005; Hoffman &
Gillam, 2004; Marton, 2008; Nickisch & von Kries, 2009). Nevertheless, in this context it is
relevant to point out that none of these studies investigated late-talking children, and that
reported sample sizes were small. Hence, the findings in the present study provide strong
evidence that differences between LTs and TDC at elementary school-age are predicted by
memory capabilities which are in turn specific to phonological processing.
Limitations
A major limitation is the (relatively) small sample of this study. Though power
analyses achieved satisfactory power for all analyses (about 95%), the sample size was
probably too small to find differences between LTs and TDC with smaller effect sizes. Thus,
reading differences between LTs and TDC cannot be definitively excluded.
Another limitation is the academic bias of our study. Though maternal education did
not differ between LTs and TDC, the maternal education in general is high in our study.
Though a high socioeconomic status (SES) of parents and children is a common concern in
academic science we cannot transfer our results to low SES families. More important,

Accepted Article
This article is protected by copyright. All rights reserved.
maternal education is only one aspect of SES while other aspects (e.g. the family income)
were not assessed. Hence, this study is not comparable to studies which are interested in the
impact of SES on language and literacy development.
The study is limited by the duration of auditory sensory memory, given that this was
investigated using a 500 ms control and a 2000 ms experimental condition. These two values
(500 ms and 2000 ms) are presumably not sensitive enough to detect continuous
modifications in the duration of auditory sensory memory. Helenius, Parviainen, Paetau, and
Salmelin (2009) presented words and pseudo-words twice to dyslexics and adults with SLI
using magnetencephalography and reported no repetition effect (decreasing N400) for
subjects with SLI. Therefore, paradigms presenting words, pseudo-words and simple non-
linguistic sounds twice with varying lengths could be more sensitive and advantageous for
future research.
Conclusions and perspectives
Even at third grade, late-talking children differ from TDC in terms of vocabulary,
verbalization of semantic relations, spelling and NWR, despite the fact that the mean of their
performance is in the normal range. Differences between the two groups resulted in the
identification of a principal component which we contend is related to phonological
processing. This component can be explained by phonological working memory capacities
while other memory scores (such as auditory sensory memory or spatial working memory)
did not account for any variance.
This study’s findings argue for the existence of persistent phonological difficulties in
late-talking children, difficulties which can be predicted by phonological working memory
capacities. Thus, we conclude that phonological processing is limited in LTs, and that this is
noticeable in early childhood in the form of reduced vocabulary size and a lack of word

Accepted Article
This article is protected by copyright. All rights reserved.
combinations. Late-talking toddlers’ language disabilities fade away in later first-language
development – a phenomenon possibly caused by numerous word repetitions in the native
language (which is overlearned and highly automatized) – though language disabilities in
early childhood could affect sustainably emotional and behavioural functioning (Clegg et al.,
2015).
Nevertheless, if lexical difficulties are persistent due to phonological working
memory limitations, second language acquisition could be impaired. Moreover, LTs’ use of
sophisticated language in academic settings could be hindered by phonological difficulties
leading to a reduction in academic opportunities for LTs. Since late-talking children differ
from TDC on a range of domains including vocabulary size of third-graders, early language
interventions targeting LTs might be important.
To date, second language acquisition, academic achievements as well as treatment
effects of early language interventions have been poorly investigated for LTs at school-age.
Thus, future research should address these issues.
References
Adams, A., & Gathercole, S. E. (2000). Limitations in working memory: Implications for
language development. International Journal of Language & Communication Disorders,
35(1), 95-116.
Archibald, L., & Gathercole, S. (2006). Visuospatial immediate memory in specific language
impairment. Journal of Speech, Language, and Hearing Research, 49(2), 265-277.
Archibald, L., & Gathercole, S. (2007). The complexities of complex memory span: Storage
and processing deficits in specific language impairment. Journal of Memory and
Language, 57(2), 177–194.
Baddeley, A. D., & Hitch, G. (1974). Working memory. Psychology of learning and
motivation, 8, 47-89.
Baddeley, A. D., & Logie, R. H. (1999). Working memory: The multiple-component model. In A
Miyake & P. Shah (Eds.), Models of working memory: Mechanisms of active maintenance
and executive control (pp. 28–61). New York: Cambridge University Press

Accepted Article
This article is protected by copyright. All rights reserved.
Baird, G., Slonims, V., Simonoff, E., & Dworzynski, K. (2011). Impairment in non-word
repetition: A marker for language impairment or reading impairment? Developmental
Medicine and Child Neurology, 53(8), 711–716.
Bakos, S., Landerl, K., Bartling, J., Schulte-Körne, G., & Moll, K. (2018).
Neurophysiological correlates of word processing deficits in isolated reading and
isolated spelling disorders. Clinical Neurophysiology, 129(3), 526-540.
Bartha-Doering, L., Deuster, D., Giordano, V., am Zehnhoff-Dinnesen, A., & Dobel, C.
(2015). A systematic review of the mismatch negativity as an index for auditory sensory
memory: From basic research to clinical and developmental perspectives.
Psychophysiology, 52(9), 1115-1130.
Bartl-Pokorny, K. D., Marschik, P. B., Sachse, S., Green, V. A., Zhang, D., Van Der Meer,
L., … Einspieler, C. (2013). Tracking development from early speech-language
acquisition to reading skills at age 13. Developmental Neurorehabilitation, 16(3), 188-
195.
Bavin, E. L., Wilson, P. H., Maruff, P., & Sleeman, F. (2005). Spatio-visual memory of
children with specific language impairment: evidence for generalized processing
problems. International Journal of Language & Communication Disorders / Royal
College of Speech & Language Therapists, 40(3), 319–32.
Beckage, N., Smith, L., & Hills, T. (2011). Small worlds and semantic network growth in
typical and late talkers. PloS One, 6(5), e19348. doi:10.1371/journal.pone.0019348
Birkel, P. (2007). WRT 3+. Weingartener Grundwortschatz Rechtschreib-Test für dritte und
vierte Klassen. Göttingen: Hogrefe.
Bishop, D. V., Holt, G., Line, E., McDonald, D., McDonald, S., & Watt, H. (2012). Parental
phonological memory contributes to prediction of outcome of late talkers from 20
months to 4 years: A longitudinal study of precursors of specific language impairment.
Journal of Neurodevelopmental Disorders, 4(1), 3. doi:10.1186/1866-1955-4-3
Bishop, D. V. M. (2007). Using mismatch negativity to study central auditory processing in
developmental language and literacy impairments: Where are we, and where should we
be going? Psychological Bulletin, 133(4), 651–672.
Bleses, D., Makransky, G., Dale, P. S., Hojen, A., & Ari, B. A. (2016). Early productive
vocabulary predicts academic achievement 10 years later. Applied Psycholinguistics, 1-
16.
Bornstein, M. H., Hahn, C. S., Putnick, D. L., & Suwalsky, J. T. (2014). Stability of core
language skill from early childhood to adolescence: A latent variable approach. Child
Development, 85(4), 1346-1356.
Botting, N., & Conti-Ramsden, G. (2001). Non-word repetition and language development in
children with specific language impairment (SLI). International Journal of Language
and Communication Disorders, 36(4), 421–432.

Accepted Article
This article is protected by copyright. All rights reserved.
Brandenburg, J., Klesczewski, J., Fischbach, A., Schuchardt, K., Büttner, G., & Hasselhorn,
M. (2015). Working memory in children with learning disabilities in reading versus
spelling: Searching for overlapping and specific cognitive factors. Journal of learning
disabilities, 48(6), 622-634.
Briscoe, J., & Rankin, P. M. (2009). Exploration of a ‘double jeopardy’hypothesis within
working memory profiles for children with specific language impairment. International
Journal of Language & Communication Disorders, 44(2), 236-250.
Clegg, J., Law, J., Rush, R., Peters, T. J., & Roulstone, S. (2015). The contribution of early
language development to children's emotional and behavioural functioning at 6 years: an
analysis of data from the Children in Focus sample from the ALSPAC birth cohort.
Journal of Child Psychology and Psychiatry, 56(1), 67-75.
Coady, J. A., & Evans, J. L. (2008). Uses and interpretations of non-word repetition tasks in
children with and without specific language impairments (SLI). International Journal of
Language Communication Disorders / Royal College of Speech Language Therapists,
43(1), 1–40.
Conti-Ramsden, G., Botting, N., & Faragher, B. (2001). Psycholinguistic markers for specific
language impairment (SLI). Journal of Child Psychology and Psychiatry, and Allied
Disciplines, 42(6), 741–48.
Cowan, N. (1995). Attention and memory: An integrated framework. New York: Oxford
University Press.
Dale, P. S., McMillan, A. J., Hayiou-Thomas, M. E., & Plomin, R. (2014). Illusory recovery:
Are recovered children with early language delay at continuing elevated risk? American
Journal of Speech-Language Pathology, 23(3), 437-447.
Dale, P. S., Price, T. S., Bishop, D. V. M., & Plomin, R. (2003). Outcomes of early language
delay: I. Predicting persistent and transient language difficulties at 3 and 4 years.
Journal of Speech, Language, and Hearing Research : JSLHR, 46(3), 544–560.
Duff, F. J., Reen, G., Plunkett, K., & Nation, K. (2015). Do infant vocabulary skills predict
school-age language and literacy outcomes? Journal of Child Psychology and
Psychiatry and Allied Disciplines, 56(8), 848–856. Edwards, S. & Reynell J. (1997).
Reynell Developmental Language Scales III. NFER-Nelson Health & Social Care.
Estes, K. G., Evans, J. L., & Else-Quest, N. M. (2007). Differences in the nonword repetition
performance of children with and without specific language impairment: A meta-
analysis. Journal of Speech, Language, and Hearing Research, 50(1), 177-195.
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G* Power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sciences. Behavior
research methods, 39(2), 175-191.
Fenson, L., Dale, P. S., & Reznick, J. S. (1993). MacArthur communicative development
inventories: User’s guide and technical manual. San Diego, CA: Singular Publishing
Group.

Accepted Article
This article is protected by copyright. All rights reserved.
Fernald, A., & Marchman, V. A. (2012). Individual differences in lexical processing at 18
months predict vocabulary growth in typically developing and late-talking toddlers.
Child Development, 83(1), 203–222.
Gathercole, S. E., & Baddeley, A. D. (1990). Phonological memory deficits in language
disordered children: Is there a causal connection? Journal of Memory and Language,
29(3), 336–360.
Girolametto, L., Wiigs, M., Smyth, R., Weitzman, E., & Pearce, P. S. (2001). Children with a
history of expressive vocabulary delay: Outcomes at 5 years of age. American Journal of
Speech-Language Pathology, 10(4), 358–369.
Grau, C., Escera, C., Yago, E., & Polo, M. D. (1998). Mismatch negativity and auditory
sensory memory evaluation: A new faster paradigm. NeuroReport, 9(11), 2451-2456.
Grimm, H., Aktaş, M., & Frevert, S. (2000). SETK-2: Sprachentwicklungstest für
zweijährige Kinder: Diagnose rezeptiver und produktiver
sprachverarbeitungsfähigkeiten. Göttingen: Hogrefe.
Grimm, H., Aktas, M. & Frevert, S. (2001). SETK 3-5. Sprachentwicklungstest für drei- bis
fünfjährige Kinder. Diagnose von Sprachverarbeitungsfähigkeiten und auditiven
Gedächtnisleistungen. Göttingen: Hogrefe.
Grimm, H., & Doil, H. (2000). Elternfragebögen für die Früherkennung von Risikokindern:
ELFRA. Göttingen: Hogrefe.
Grossheinrich, N., Kademann, S., Bruder, J., Bartling, J., & Von Suchodoletz, W. (2010).
Auditory sensory memory and language abilities in former late talkers: A mismatch
negativity study. Psychophysiology, 47(5), 822–830.
Hammer, C. S., Morgan, P., Farkas, G., Hillemeier, M., Bitetti, D., & Maczuga, S. (2017).
Late talkers: A population-based study of risk factors and school readiness
consequences. Journal of Speech, Language, and Hearing Research, 60(3), 607-626.
Helenius, P., Parviainen, T., Paetau, R., & Salmelin, R. (2009). Neural processing of spoken
words in specific language impairment and dyslexia. Brain, 132(7), 1918–1927.
Hellbrügge, T., Lajosi, F., Menara, D., Schamberger, R. & Rautenstrauch, T. (1985):
Münchener Funktionelle Entwicklungsdiagnostik. Lübeck: Hansisches Verlagskontor.
Henry, L. A., & Botting, N. (2017). Working memory and developmental language
impairments. Child Language Teaching and Therapy, 33(1), 19-32.
Hick, R., Botting, N., & Conti-Ramsden, G. (2005). Cognitive abilities in children with
specific language impairment: Consideration of visuo-spatial skills. International
Journal of Language & Communication Disorders / Royal College of Speech &
Language Therapists, 40(2), 137–49.

Accepted Article
This article is protected by copyright. All rights reserved.
Hoffman, L. M., & Gillam, R. B. (2004). Verbal and spatial information processing
constraints in children with specific language impairment. Journal of Speech, Language,
and Hearing Research : JSLHR, 47(1), 114–25.
Landerl, K., Ramus, F., Moll, K., Lyytinen, H., Leppänen, P. H., Lohvansuu, K., ... &
Schulte-Körne, G. (2013). Predictors of developmental dyslexia in European
orthographies with varying complexity. Journal of Child Psychology and
Psychiatry, 54(6), 686-694.
Law, J., Garrett, Z., & Nye, C. (2004). The efficacy of treatment for children with
developmental speech and language delay/disorder: A meta-analysis. Journal of Speech,
Language, and Hearing Research, 47(4), 924-943.
Lee, J. (2011). Size matters: Early vocabulary as a predictor of language and literacy
competence. Applied Psycholinguistics, 32(1), 69-92.
Lenhard, W., & Schneider, W. (2006). ELFE 1-6. Ein Leseverständnistest für Erst-bis
Sechsklässler. Göttingen: Hogrefe.
Lukács, Á., Ladányi, E., Fazekas, K., & Kemény, F. (2016). Executive functions and the
contribution of short-term memory span in children with specific language impairment.
Neuropsychology, 30(3), 296-303.
Lum, J. A. G., Conti-Ramsden, G., Page, D., & Ullman, M. T. (2012). Working, declarative
and procedural memory in specific language impairment. Cortex; a Journal Devoted to
the Study of the Nervous System and Behavior, 48(9), 1138–54.
Maekawa, J., & Storkel, H. L. (2006). Individual differences in the influence of phonological
characteristics on expressive vocabulary development by young children. Journal of
Child Language, 33(3), 439–59.
Marschik, P. B., Einspieler, C., Garzarolli, B., & Prechtl, H. F. (2007). Events at early
development: Are they associated with early word production and neurodevelopmental
abilities at the preschool age? Early Human Development, 83(2), 107-114.
Marton, K. (2008). Visuo-spatial processing and executive functions in children with specific
language impairment. International Journal of Language & Communication Disorders /
Royal College of Speech & Language Therapists, 43(2), 181–200.
Melchers, P. & Preuß, U. (2009). Kaufman Assessment Battery for Children, dt. Version (K-
ABC). Frankfurt am Main: Pearson Assessment.
Moll, K., & Landerl, K. (2010). SLRT-II: Lese-und Rechtschreibtest; Weiterentwicklung des
Salzburger Lese-und Rechtschreibtests (SLRT). Bern: Huber.
Moyle, J., Stokes, S. F., & Klee, T. (2013). Early language delay and specific language
impairment. Developmental Disabilities Research Review, 17(2),160-169.

Accepted Article
This article is protected by copyright. All rights reserved.
Moyle, M. J., Weismer, S. E., Evans, J. L., & Lindstrom, M. J. (2007). Longitudinal
relationships between lexical and grammatical development in typical and late-talking
children. Journal of Speech Language and Hearing Research, 50, 508–528.
Näätänen, R., Jacobsen, T., & Winkler, I. (2005). Memory-based or afferent processes in
mismatch negativity (MMN): A review of the evidence. Psychophysiology, 42(1), 25-32.
Newcomer, P.L., & Hammill, D.D. (2008). Test of Language Development (TOLD-P: 4).
Austin: Pro-Ed.
Nickisch, A., & von Kries, R. (2009). Short-term memory (STM) constraints in children with
specific language impairment (SLI): Are there differences between receptive and
expressive SLI? Journal of Speech, Language, and Hearing Research: JSLHR, 52(3),
578–95.
Paul, R., Murray, C., Clancy, K., & Andrews, D. (1997). Reading and metaphonological
outcomes in late talkers. Journal of Speech, Language, and Hearing Research : JSLHR,
40(5), 1037–1047.
Petermann, F., Metz, D., & Fröhlich, L. P. (2010). SET 5-10: Sprachstandserhebungstest für
Kinder im Alter zwischen 5 und 10 Jahren. Göttingen: Hogrefe.
Reilly, S., Wake, M., Ukoumunne, O. C., Bavin, E., Prior, M., Cini, E., … Bretherton, L.
(2010). Predicting language outcomes at 4 years of age: Findings from Early Language
in Victoria Study. Pediatrics, 126(6), e1530–e1537.
Rescorla, L. (2002). Language and reading outcomes to age 9 in late-talking toddlers. Journal
of Speech, Language and Hearing Research, 45(2), 360–71.
Rescorla, L. (2005). Age 13 language and reading outcomes in late-talking toddlers. Journal
of Speech, Language, and Hearing Research, 48(2), 459–472.
Rescorla, L. (2011). Late talkers: Do good predictors of outcome exist? Developmental
Disabilities Research Reviews, 17(2), 141–50. Rescorla, L., Dahlsgaard, K., & Roberts,
J. (2000). Late-talking toddlers: MLU and IPSyn outcomes at 3;0 and 4;0. Journal of
Child Language, 27(3), 643–664.
Rescorla, L., & Turner, H. L. (2015). Morphology and Syntax in Late Talkers at Age 5.
Journal of Speech, Language, and Hearing Research, 58(2), 434-444.
Rice, M. L., Taylor, C. L., & Zubrick, S. R. (2008). Language outcomes of 7-year-old
children with or without a history of late language emergence at 24 months. Journal of
Speech, Language, and Hearing Research : JSLHR, 51(2), 394–407.
Sachse, S., & Von Suchodoletz, W. (2008). Early identification of language delay by direct
language assessment or parent report? Journal of Developmental and Behavioral
Pediatrics : JDBP, 29(1), 34–41.
Shankweiler, D., & Crain, S. (1986). Language mechanisms and reading disorder: A modular
approach. Cognition, 24, 139–168.

Accepted Article
This article is protected by copyright. All rights reserved.
Stokes, S. F., & Klee, T. (2009). Factors that influence vocabulary development in two-year-
old children. The Journal of Child Psychology and Psychiatry and Allied Disciplines,
50(4), 498–505.
Swanson, H. L., & Ashbaker, M. H. (2000). Working memory, short-term memory, speech
rate, word recognition and reading comprehension in learning disabled readers: does the
executive system have a role? Intelligence, 28(1), 1-30.
Szagun, G., Steinbrink, C., Franik, M., & Stumper, B. (2006). Development of vocabulary
and grammar in young German-speaking children assessed with a German language
development inventory. First Language, 26(3), 259-280.
Tellegen, P. J., Winkel, M., Wijnberg-Williams, B. J. & Laros, J. A. (1998): Snijders-
Oomen Non-verbaler Intelligenztest - SON-R 2 1/2 - 7. Frankfurt: Swets & Zeitlinger.
Thal, D. J., Miller, S., Carlson, J., & Vega, M. M. (2005). Nonword repetition and language
development in 4-year-old children with and without a history of early language delay.
Journal of Speech, Language and Hearing Research, 48(6), 1481–1495.
Vugs, B., Cuperus, J., Hendriks, M., & Verhoeven, L. (2013). Visuospatial working memory
in specific language impairment: A meta-analysis. Research in Developmental
Disabilities, 34(9), 2586-2597.
Weismer, S., Tomblin, J. B., Zhang, X., Buckwalter, P., Chynoweth, J. G., & Jones, M.
(2000). Nonword repetition performance in school-age children with and without
language impairment. Journal of Speech Language and Hearing Research, 43, 865–878.
Weismer, S. E., Venker, C. E., Evans, J. L., & Moyle, M. J. (2013). Fast Mapping in Late-
Talking Toddlers. Applied Psycholinguistics, 34(1), 69–89.
Figure legends
Figure 1: Oddball paradigm (upper panel: 500 ms condition; lower panel: 2000 ms condition)
Figure 2: Mismatch Negativity (MMN) for the experimental condition (ISI: 2000 ms) and the
control condition (ISI. 500 ms) depicting late-talking (black lines) and typical
developing children (grey lines) splitted by the mean in above-average (solid lines)
and below-average performance (dashed lines)

Accepted Article
This article is protected by copyright. All rights reserved.
Table 1: Characteristics of LTs and TDC
LTs (N = 39)
Controls (N = 39)
N
fisher’s
exact test
p
Boys/Girls (N)
24/15
22/17
.82
Handedness (N)
(right/left)
32/7
37/2
.15
Rank data
U
p
Maternal educationa
(median/range)
no graduation (1)
special needs education (2);
general school curriculum (3);
intermediate education (4);
higher education (5)
5/2
0 %
0 %
15 %
31 %
54 %
5/2
0 %
0 %
13 %
23 %
64 %
-.85
.39
Raw data (M, SD)
T
p
Age (in month)
Month of education
112 + 3.78
26.23 + 1.13
109 + 3.64
25.97 + 0.84
-3,57
-1.13
< .01
.26
IQ (M, SD)
T
p
Non-verbal intelligenceb
104.92 + 10.49
108.24 + 9.85
1.44
.15
a no graduation (1); education for children with special needs (2); general school curriculum (3); the intermediate
school curriculum (4); higher education entrance diploma; or Abitur (5)
b four subtests of the Snijders-Oomen non-verbal-intelligence test (Tellegan et al., 1998)

Accepted Article
This article is protected by copyright. All rights reserved.
Table 2: MANOVA of language, reading and spelling at the third grade for late talkers
(LTs; N = 39) and typical developing children (TDC; N = 39)
z-values (N = 78, df = 1)
Variable
LT
TDC
F
MSE
p
Partial
Eta Squared
Language
Vocabulary
.7 (1.0)
1.2 (1.0)
5.70
5.18
.02
.07
Semantic relations
.6 (.9)
1.3 (1.0)
10.41
8.87
<.001
.12
Language comprehension I
(comprehension of a text)
2.1
(1.4)
2.1 (1.5)
0.01
0.01
.94
Language comprehension II
(syntactic understanding)
.9 (1.6)
1.2 (1.6)
0.60
1.52
.45
Grammar I (inflections –
plurals of real and
pseudowords)
.5 (.8)
.8 (.8)
3.00
1.97
.09
Grammar II (correcting
incorrect sentences)
1.3
(1.6)
1.8 (1.4)
1.99
4.68
.16
Reading/
Spelling
Reading comprehension
.1 (.8)
.3 (.8)
2.45
1.51
.12
Reading speed
-.3 (.7)
-.1 (.8)
1.57
0.92
.21
Spelling
-.01
(1.0)
.6 (1.1)
6.22
7.20
.02
.08
Speed
-.2 (.8)
.1 (1.0)
1.18
0.97
.28
Memory
Phonological memory
(NWR)*
15.5
(2.8)
17.3 (2.4)
10.78
70.20
<.01
.12
* Raw values, NWR = Non-word repetition

Accepted Article
This article is protected by copyright. All rights reserved.
Table 3: Final model predicting differences in achievements between late talkers and
typically developing children at third grade.
Final model: R2 = .30
Note: Intervention was coded dichotomously (children who received intervention: 31% vs.
those who did not: 69%)
β
T
p
Predictors
Non-word
repetition
.49
4.66
< .01
Word order
.24
2.03
.04
Phonological
awareness
.22
1.73
.08
Covariates
Sex
.01
.07
.95
Intervention
.16
1.4
.17
IQ
.19
1.65
.10

Accepted Article
This article is protected by copyright. All rights reserved.

Accepted Article
This article is protected by copyright. All rights reserved.