ORIGINAL RESEARCH ARTICLE

published: 06 June 2014

HUMAN NEUROSCIENCE doi: 10.3389/fnhum.2014.00386

Neurocognitive profiles of learning disabled children with

neurofibromatosis type 1
Miladys Orraca-Castillo 1 , Nancy Estévez-Pérez 2* and Vivian Reigosa-Crespo 2
1
Provincial Centre for Medical Genetics, Pinar del Río, Cuba
2
Department of Developmental Cognitive Neuroscience, Cuban Centre for Neuroscience, Havana, Cuba

Edited by: Neurofibromatosis 1 (NF1) is a genetic condition generally associated with intellectual
John J. Foxe, Albert Einstein College deficiency and learning disabilities. Although there have been groundbreaking advances
of Medicine, USA
in the understanding of the molecular, cellular, and neural systems underlying learning
Reviewed by:
deficits associated to NF1 in animal models, much remains to be learned about the
Natalie Russo, City College of New
York, USA spectrum of neurocognitive phenotype associated with the NF1 clinical syndrome. In
Bonita P. Klein-Tasman, University of the present study, 32 children with NF1 ranging from 7 to 14 years were evaluated with
Wisconsin - Milwaukee, USA neurocognitive tests dedicated to assess basic capacities which are involved in reading and
*Correspondence: mathematical achievement. Deficits in lexical and phonological strategies and poor number
Nancy Estévez-Pérez, Nancy
facts retrieval were found underlying reading and arithmetic disorders, respectively.
Estévez-Pérez, Department of
Developmental Cognitive Additionally, efficiencies in lexical/phonological strategies and mental arithmetic were
Neuroscience, Cuban Centre for significant predictors of individual differences in reading attainment and math. However,
Neuroscience, Ave. 25, No.15202 deficits in core numeric capacities were not found in the sample, suggesting that it is
esq. 158, Cubanacán, Playa, Código
Postal 11600, Apartado 6412,
not responsible for calculation dysfluency. The estimated prevalence of Developmental
Habana 6414, Cuba Dyscalculia was 18.8%, and the male:female ratio was 5:1. On the other hand, the
e-mail: nancy.estevez@ prevalence of Developmental Dyslexia was almost 3 times as high (50%), and no gender
cneuro.edu.cu differences were found (male: female ratio = 1:1). This study offers new evidence to
the neurocognitive phenotype of NF1 contributing to an in depth understanding of this
condition, but also to possible treatments for the cognitive deficits associated with NF1.
Keywords: neurofibromatosis type 1, specific learning disabilities, developmental dyscalculia, developmental
dyslexia, phonological dyslexia, surface dyslexia, arithmetic dysfluency, gender ratio

INTRODUCTION with NF1 probably reflects the relative contributions of multi-

Neurofibromatosis 1 (NF1) is an autosomal dominant disease ple cellular and molecular defects (Diggs-Andrews and Gutmann,
caused by mutations of the NF1 gene, a tumor suppressor gene, 2013). Considering this, it has been suggested to deal with the
on chromosome 17. Almost all affected individuals exhibit multi- cognitive deficits in children with NF1 by integrating therapies
ple cafe’au lait spots and cutaneous neurofibromas. Less frequent specifically designed to target the biochemical and neurochem-
but more serious manifestations include central nervous system ical abnormalities unique to each child (Diggs-Andrews and
gliomas, plexiform, neurofibromas, and malignant peripheral Gutmann, 2013).
nerve sheath tumors (Rasmussen and Friedman, 2013). Children with NF1 usually exhibit low intellectual capaci-
Learning, cognitive, and neurobehavioral deficits are highly ties. Attention Deficit and/or Hyperactivity Disorders (ADHD),
prevalent in this condition (Acosta et al., 2012). As a single gene depression, poor school attainment and postschool outcomes are
disorder, NF1 provides a unique genetic model to identify and frequently reported as well (Mazzocco, 2001; Billingsley et al.,
analyze the molecular and cellular bases underlying cognitive dys- 2004; Payne et al., 2010; Rasmussen and Friedman, 2013). The
function. Thus, this disorder has received the full attention of cognitive impairments more frequently described include spe-
the scientific community dedicated to the translation of basic cific learning disabilities (SLD)—such as Developmental Dyslexia
or bench science, to bedside clinical practice or dissemination (DL) and Developmental Dyscalculia (DD)—visuo-spatial orien-
to population-based community interventions (translational sci- tation, and attention deficits, language acquisition, spatial mem-
ence) (Acosta et al., 2012). ory and executive functions defects (Billingsley et al., 2004; Payne
Nevertheless, the marked clinical variability described in the et al., 2010).
syndrome suggests that NF1 is composed of numerous distinct There are various genetic syndromes which exhibit SLD.
diseases, each defined by factors including patient age, patient Together with NF1, Turner, Fragile X, and Velocardiofacial syn-
sex, the timing of NF1 inactivation, the specific cell type, genomic dromes are the most studied (Mazzocco, 2001; Molko et al.,
modifiers, and microenvironmental influences. No single molec- 2004; Murphy and Mazzocco, 2008). SLD have been reported
ular abnormality seems to underlie all of the cognitive dysfunc- in 30–70% of children with NF1 (Hyman et al., 2005; Acosta,
tion observed in children with NF1. Rather, the specific collection 2007). Nevertheless, the diagnosis of SLD has mainly focused
of cognitive and behavioral deficits exhibited by any given child on IQ-achievement discrepancy (Hyman et al., 2005; Watt et al.,

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 1

Orraca-Castillo et al. Learning disabilities in NF1

2008). The discrepancy model has been criticized for numerous neurocognitive phenotype of NF1 syndrome would contribute to
authors. Over the past 30 years dozens of research articles have the development of specific clinical and behavioral interventions
provided empirical evidence of the problems inherent to the IQ- for NF1 syndrome. In addition it would contribute to the mod-
achievement discrepancy model (Restori et al., 2009). It is now eling of the core cognitive and neurobiological foundations of
clear that establishing the discrepancy between intelligence and typical and atypical learning of reading and mathematics.
achievement is not particularly useful either for assessment or
intervention purposes.
MATERIALS AND METHODS
ETHICS STATEMENT
As an alternative point of view, current theories of typi-
cal cognitive development postulate that achievement is based The study was approved by the Ethics Committee of the
on a restricted set of core systems, defined as domain-specific Provincial Center for Medical Genetics (PCMG), Pinar del Río,
representational primitives that lead and constrain the cultural and by the Institutional Review Board of the Cuban Center for
learning (Spelke and Kinzler, 2007). Thus, the focus of attention Neuroscience and it’s in line with the Cuban Science, Technology
has shifted from higher level, school-like arithmetic and read- and Environment Ministry’s approved projects. No identity
ing skills to an analysis of basic capacities, or domain-specific revealing photographs were taken. Written consent was obtained
cognitive skills. from all parents, and all participants provided verbal assent for all
According to this approach, DD is considered a congenital assessments.
and persistent disability in achieving normal levels of arithmetical
PARTICIPANTS
skills that arise when the specialized number system (Feigenson
Thirty two (32) children (14 girls) with NF1 registered and
et al., 2004; Butterworth, 2010) fails to develop normally with cor-
screened by the PCMG in the Pinar del Río province, ranging
responding deleterious effects in the acquisition of higher level
from 7 to 14 years of age were included [7–8 years old: 5 chil-
math skills. Neuroimaging studies of neural foundations of DD
dren (3 girls); 9–10 years old: 7 children (4 girls); 11–12 years old:
compared to controls describe low number-related activations in
10 children (2 girls); 13–14 years old: 10 children (5 girls)]. NF1
the Intraparietal Sulcus (an area dedicated to number processing)
was diagnosed according to the criteria described by the National
and the recruitment of more distributed brain regions (possibly
Institutes of Health Consensus Development Conference (Ferner
reflecting compensatory strategies) (Kaufmann et al., 2011).
et al., 2007). Among the main clinical signs, children included in
Although this approach focuses on core numeric capacities, it
the sample exhibited (more than six) pigmented birthmarks typ-
also accepts the contribution of other domain-general cognitive
ically distributed around the torso and the arms/legs, freckling
processes such as executive functions to arithmetical attainment
of the armpits, skeletal abnormalities and (two or more) Lisch
(Butterworth, 2010). A deficient recruitment of frontal brain
nodules and/or neurofibromas.
regions and supramarginal and postcentral gyrus, areas found to
support domain-general processing in typically developing chil- TASKS
dren (Kaufmann et al., 2011) supports the role of domain-general Attainment tests
cognitive processes in the numerical cognition. Two attainment tests were used in order to detect numerical and
On the other hand, DL is considered a congenital and persis- reading impairments in the classroom. The MAT (Mathematics
tent disability in reading and comprehension, and exhibits deficits Attainment Test) and RCAT (Reading and Comprehension
in phonological processing - the awareness of the sound struc- Attainment test) are non-standardized curriculum-based mea-
ture of words—and, in some cases, a more fundamental deficit surements developed by researchers at the Ministry of Education
in rapid auditory processing (Temple, 2002). Neuroimaging stud- (Bernabeu and León, 2003, Pers. Commun.) and employed
ies of neural foundations of DL describe underactivation in the throughout Cuban schools. The authors created the measures
left hemisphere inferior frontal gyrus and temporoparietal cor- by selecting problem types and texts representing a propor-
tex (underlying deficient phonological processing) and the visual tional sampling of the computation and reading skills within the
word form area in the left hemisphere occipitotemporal region national curriculum.
(important in word recognition) (Richlan, 2012). Inferior frontal MAT comprised eight computational problems by each respec-
white matter decreases in the left frontal and parietal portions tive grade (2 to 9th). Total score was up to 8 (one per each problem
of the arcuate fasciculus and other left perisilvian areas have also performed correctly). Children were considered to fail the MAT
been associated to the disorder (Silani et al., 2005). test if obtaining a total score ≤5.
To the best of our knowledge, the identification of DD focused RCAT comprised a text and five comprehension questions
on defective basic numerical capacities and math fluency has not related to it. Total score was up to 5 (one per each correct
been previously addressed in children with NF1 and only one answer). Children were considered to fail RCAT if obtaining a
study evaluated DL subtypes considering phonological and lexi- total score ≤3.
cal deficits in this genetic disorder (Watt et al., 2008). The present
study aims to evaluate a group of children with NF1 in order Core cognitive assessment
to determine the presence of DD and DL based on profiles of Item-timed computerized tasks from two standardized neurocog-
defective core capacities, using neurocognitive tests specifically nitive batteries designed in order to evaluate core cognitive pro-
designed and standardized for the diagnosis of these SLD. Gender cesses recruited by numerical processing and reading were used:
ratios of the detected disorders are also examined in the sample mental arithmetic, dot enumeration, number magnitude com-
studied. The confirmation of the presence of the disorders in the parison and simple reaction time tasks from the Basic Numerical

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 2

Orraca-Castillo et al. Learning disabilities in NF1

Battery (BNB) (Reigosa-Crespo et al., 2012) and the word and trials were presented. Five practice trials were given before starting
pseudoword reading task from Batería de Trastornos de la Lectura the test. Reaction times (RT) were recorded. The simple reaction
(BTL) (Reigosa-Crespo et al., 1994). time test was not analyzed by itself. It was considered a baseline
measure of processing speed. The reaction times on all numeri-
Standardization data of BNB. BNB was standardized using a cal tasks described below were adjusted taking this measure into
stratified random sampling strategy (Pedhazur and Pedhazur- account.
Schmelkin, 1991) which reduces sampling variability by creating
relatively homogeneous strata with respect to the dependent vari- Mental arithmetic. Fifteen simple additions, 15 subtractions and
able of interest. This strategy allows the test developer to produce 15 multiplications were presented in three separate blocks. All
norms with less sampling error as would a simple randomized involved single-digit numbers from 2 to 9. Items were presented
sample (Crocker and Algina, 1986). on the computer screen (in the form “2+4”). No ties (e.g., 3 + 3,
The school-age population (n = 11, 652) of a municipality of 5 × 5) were presented, and items were not repeated. Two prac-
Havana was first split up by grade (2 to 9th) and then, divided into tice trials were given before the start of each block. Children were
nine strata on the basis of the score obtained on MAT test (0–8). asked to type in the answer as quickly as they could without mak-
Each stratum was initially treated independently. Thus, children ing any mistakes. Reaction time was measured with the first key
within each stratum were randomly selected and individual esti- stroke. Hits were also recorded.
mates (proportions) were obtained. These estimates were then
weighted to arrive at an estimate for the population parameters. Dot enumeration. Randomly arranged dots ranging from 1 to 9
The normative sample was comprised by 376 children (188 were presented on the computer display. Children were asked to
boys and 188 girls) homogeneously distributed by grade. No estimate the number of dots and to respond as quickly as they
significant differences were found between the proportions of could without making mistakes. Eighteen trials were presented
individuals randomly selected and the proportions of individuals altogether, with each number from 1 to 9 being presented twice
of general population within each stratum. in a pseudo-random order with the proviso that no item occurred
Medians of reaction time for correct responses in the numer- twice in succession. Five practices were given before starting the
ical tasks were adjusted subtracting each by the median of the test. Reaction times and hits were recorded.
simple RT for that participant (adjRTs). Then, an efficiency mea-
sure (EM) for each task was calculated using adjRTs divided by the Number Magnitude Comparison. Children were presented with 25
proportion of hits (EM = adjRT/Hits).This is an inverse measure, pairs of digits (numerosities: 1–9, numerical distances: 1–3) on
so higher scores represent worse performance. Since EMs did not the computer screen. The children were asked to compare the
fit a normal distribution (the data were skewed to the right), numbers on left-right direction (e.g., 5 < 7, 7 > 5); accordingly,
a logarithmic transformation of EMs (logEMs) was performed the number in the left side appeared before the number in the
(Reigosa-Crespo et al., 2012). right side. Both numbers remained on the screen until the answer
The norms for the BNB tasks were calculated using the General was recorded. Five practices were given before starting the test.
Regression Model module of Statistica 6.1 software. A regression Reaction Times were recorded by pressing the key corresponding
of logEMs as function of age with a linear and a quadratic compo- to the answer (1 for “smaller than,” 2 for “bigger than” and 3 for
nent was estimated for each task. The residuals of the regressions “equal as”). Hits were also recorded.
were obtained for each individual. The normative parameters for EM for the three numerical tasks were produced as explained
all the tasks were obtained and the individual residual for each when describing BNB sstandardization procedure. For the men-
test obtained from a given child are used to calculate a Z-score tal arithmetic test, the EM scores for each operation (addition,
by test, thus allowing the comparison of new subjects with the subtraction, and multiplication) were calculated for each child.
corresponding (age-appropriate) normative data. The mean of these medians for each child was then used as a
measure of efficiency on the mental arithmetic test overall. As
Tests description in a previous study [11], these two measures (RT and propor-
Basic Numerical Battery (BNB). BNB (Reigosa-Crespo et al., tion of hits) were used because it had been noted that children
2012) is a battery of item-timed computerized tests. BNB includes with low numeracy tend to adopt strategies that produce gen-
two numerical capacity tests: dot enumeration and number mag- erally accurate answers but extremely long RT latencies; or they
nitude comparison, and also, a test of mental arithmetic fluency. would simply guess quickly, leading to inaccurate answers but
Each test included practice trials for ensuring the understanding short RT latencies. Note that higher EM scores represent worse
of the instructions. The children always had to give a response performance.
by pressing the corresponding key (thus misses were not measur- Individual Z-score for each test was calculated using the mean
able). Only the keys of the numeric pad (right side of the key- and standard deviation (SD) of the residuals of the regressions of
board) were available for response (excepting the Simple Reaction EMs as a function of age, estimated from the normative sample.
Time task).
Word and pseudoword reading. This task evaluates the phono-
Simple reaction time. Children were asked to press the space key logical and lexical strategies involved in written word decoding
as soon as they saw a square in the center of display. The inter- and is included in the standardized BTL battery (Reigosa-Crespo
stimulus presentation time was variable (500–1500 ms). Twenty et al., 1994). The authors reported the word and pseudoword

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 3

Orraca-Castillo et al. Learning disabilities in NF1

reading task exhibited appropriate validity and reliability values was deficient (assessed with mental arithmetic task), and core
(Reigosa et al., 2002). numerically disabled when only their estimation and/or num-
The children were required to read 30 words balanced by fre- ber comparison capacities where deficient (assessed with the
quency, number of letters and syllables and 30 pseudowords. corresponding numerical core cognitive tasks).
Each stimulus was presented sequentially. The trial ended after
the child responds or after 5000 ms with no response. Ten Learning disabilities associated with intellectual disability. It
practice trials were presented initially to ensure that the chil- was determined by failure in RCAT test, Z-score < 2 SDs in the
dren understood the task. Responses were verbal and triggered word and pseudo-word reading task and/or failure in the MAT
a voice activated key which measured vocal response latencies test, and/or Z-score < 2 SDs in at least one of the three capac-
from the onset of presentation. Errors were recorded by the ity tests (dot enumeration, numerical magnitude comparison
experimenter. and mental arithmetic), associated with low intellectual capacity
Median vocal response times (vRT) for correct responses for (≤5th percentile in RCPM test).
word and pseudo-words were calculated separately. Then the
correponding efficiency measure (EM) was calculated by diving No Learning Disabilities. Children with normal academic
the vRT by the proportion of hits (EM = vRT/Hits). As in the achievement and intellectual capacities and Z-scores
BNB tasks, individual Z-scores for words and pseudo-words were between ± 2 SD.
calculated using the mean and standard deviation (SD) of the We also tested whether efficiency measures in numerical pro-
residuals of the regressions of EMs as a function of age, estimated cessing (efficiency for enumeration and number comparison),
from the normative sample. arithmetic fluency (efficiency for mental arithmetic) and reading
(efficiency for reading words and pseudowords) predicted indi-
Intellectual capacity vidual variations in math and reading attainment. Generalized
Raven’s Colored Progressive Matrices (RCPM) test (Raven et al., linear models were performed for each predictor. In the mod-
1992). els, EMs were defined as continuous predictor variables. MAT
This test was administered as a measure of non-verbal reason- scores (0–8) and RCAT scores (0–5) were defined separately as
ing ability. In this test, a colored pattern is shown with a missing dependent variables. The models assumed an ordinal multino-
piece. Below the pattern, six pieces, all fitting in the blank but with mial distribution of dependent variables because they may be
different patterns are shown. The child has to select the piece that ordered as categories.
fits in the pattern above. The total number of correct selections Male:female (M:F) ratios corresponding to each classification
was recorded. Each child completed the entire test, consisting of were calculated.
36 items.

Procedure RESULTS
All children were evaluated at the PCMG. The individual results Learning disabilities were diagnosed in 17 (53.1%) of the 32 chil-
were compared with standardized values of each test. According dren evaluated (Table 1). Co-morbidities between DD and DL
to the diagnostic criteria for specific learning disorders proposed were observed in the sample; nevertheless double dissociations
in the Diagnostic and Statistical Manual of Mental Disorders between DD and DL were also detected. Ideally, a double dissocia-
(American Psychological Association, 1994), children were clas- tion is evinced when at least one case exhibits intact process A and
sified as follows: impaired process B, and at least another case exhibits impaired
process A and intact process B (Caramazza, 1986; Shallice, 1988).
Developmental dyslexia. It was determined by failure in both, These complementary patterns of deficits, in addition, suggest
RCAT test and word and/or pseudoword reading test (a Z-score independent neural networks subserving the processes. Eleven
< 2 SDs), and normal intellectual capacity (score greater that the children exhibited DL without any deficiency in number process-
5th percentile in the RCPM test). ing and calculation, and one child showed the opposite pattern:
Concerning DL neurocognitive profile, children were labeled DD with spared reading skills.
as phonological dyslexics when only their phonological read- Among the children who showed only one SLD, different neu-
ing skills (assessed with pseudoword reading) were impaired, rocognitive profiles were also observed (Tables 2, 3). Seven of
and surface dyslexics when only their orthographic reading skills the 11 children diagnosed with DL, presented the mixed sub-
(assessed with high-frequency word reading) were impaired; type and four, the surface (dyslexical) subtype. Interestingly,
when both were deficient, they were said to have a mixed pro- children with phonological deficits and spared lexical skills
file. The cutoff for defining a reading skill as impaired was 2 SD were not found. The NF1 children with DD, exhibited selec-
below the mean of the normative group. tive deficit in mental calculation (arithmetic dysfluency) asso-
ciated with spared enumeration and magnitude comparison
Developmental dyscalculia. It was determined by failure in MAT skills.
test and at least, in one of the BNC tests (a Z-score < 2 SDs Additionally, EM of mental arithmetic were significant pre-
in the dot enumeration, magnitude comparison and/or mental dictors of MAT scores [W(1, 8) = 4.04, p < 0.05]. EMs of reading
calculation tasks) and normal intellectual capacity. words and pseudowords were also significant predictors of RCAT
Concerning DD neurocognitive profile, children were labeled scores [W(1, 4) = 9.31, p < 0.01 and W(1, 4) = 3.98, p < 0.05
as arithmetically dysfluent when their mental arithmetic fluency respectively].

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 4

Orraca-Castillo et al. Learning disabilities in NF1

Gender ratios suggest that boys and girls had similar probabil- 18.8% of the NF1 sample. Finally, nine of the NF1 children
ities of having DL. In contrast, boys were 5 times more likely to (28.1% of the sample) did not show learning disorders.
suffer DD than girls.
Another finding of the study is the detection of generalized DISCUSSION
learning disabilities associated with low intellectual capacity in The prevalence of SLD in the sample of children NF1 is in accor-
dance with that reported in international publications, which
ranges between 30 and 65% of the affected children (North et al.,
Table 1 | Prevalence of learning disabilities in the sample of children 1997; Hyman et al., 2005; Ferner et al., 2007). Another study,
with NF1. including cognitive assessments of basic capacities also reported
Classification N Prevalence* (%) M:F ratio higher proportion of reading, spelling, and mathematics deficits
(51%) in children with NF1 in contrast with specific learning dis-
Specific learning disabilities 17 53.1 1.18:1 abilities defined by IQ–achievement discrepancies (only 20% of
DL 16 50.0 1:1 the children). Co-morbidities between Developmental Dyslexia
DD 6 18.8 5:1 and Dyscalculia were also observed in the sample, in accordance
Learning disabilities associated to 6 18.8 1:1 with previous reports (Mazzocco, 2001).
intellectual disability In general, the studies dedicated to the evaluation of core
No learning disabilities 9 28.1 2:1 numeric processes and reading skills in genetic syndrome pop-
ulations are scarce (Mazzocco, 2001; Bruandet et al., 2004; Molko
DL, Developmental Dyslexia; DD, Developmental Dyscalculia; N, number of chil-
dren; M:F ratio, male/female ratio.
et al., 2004). Most of the current research is focused on the use
*Clinical prevalence calculated based on sample size (32 children). of academic achievement tests and the observation of a vari-
ety of general and specific cognitive deficits, usually disregarding
the modular systems involved in cognitive processing; which
Table 2 | Distribution of learning disabled children with NF1 according are increasingly becoming relevant in neurocognitive models of
to the neurocognitive profile. typical and atypical learning (Spelke and Kinzler, 2007).
The unvailability of detailed reports of learning disabilities
Classification N (%)
diagnosed in considering the presence of specific deficits in
DL (Mixed subtype) 7 21.9 numeric and arithmetic skills in NF1 does not allow the direct
DL (Surface subtype) 4 12.5 comparison of the results obtained in this study with those of
DD (arithmetic dysfluency) 1 3.1 the literature (Mazzocco, 2001; Butterworth and Reigosa-Crespo,
DL (Mixed subtype)/DD (arithmetic dysfluency) 3 9.4 2007; Clements-Stephensa et al., 2008). Nevertheless, the preva-
DL (Surface subtype)/DD (arithmetic dysfluency) 2 6.2 lence estimates based on math achievement measures in the gen-
eral population range from 3 to 14% (median of approximately
Total 17 53.1 6%), (Kosc, 1974; Badian, 1983; Lewis et al., 1994; Gross-Tsur

Table 3 | Cognitive measures in the sample of children with NF1 according to the neurocognitive profile.

Measures No LD (N = 9) DL (N = 11) DL DD (N = 5) LD & ID (N = 6)

Mean SD Mean SD Mean SD Mean SD

RT 496.111 204.172 481.818 157.6 453.400 208.211 392.667 122.463

Age 9.667 1.732 11.636 2.111 11.4 1.517 12.5 1.975
RCPM test+ 22.556 6.346 23.75 3.536 25.4 4.722 13.333 2.338
RCAT+ 4.111 0.928 2.909 1.3 2.4 1.14 2.167 1.472
MAT+ 6.625 1.768 3.7 1.829 4.0 1.826 3.0 2.449
Enumeration* 2537.906 705.081 2365.341 979.122 3041.657 389.225 2612.081 344.765
Number Comparison* 1795.130 390.043 1682.463 435.104 2372.540 1212.735 1988.895 637.409
Mental Arithmetic* 3621.818 1511.766 5077.210 4650.242 8544.266 1346.224 5539.682 4096.582
Word reading* 2018.754 276.394 2530.147 578.286 2939.580 423.303 3030.208 1298.428
Pseudoword reading* 2858.331 397.089 4755.116 1774.514 4858.311 1839.970 4335.839 1278.204

+ Raw scores.
*Efficiency Measures.
DL: Developmental Dyslexia.
DL DD: Developmental Dyslexia and Developmental Dyscalculia.
LD & ID: Learning Disabilities associated to intellectual disability.
No LD: No Learning Disabilities.
SD: Standard Deviation.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 5

Orraca-Castillo et al. Learning disabilities in NF1

et al., 1996; Desoete et al., 2004). The prevalence of DD found the dorsolateral prefrontal cortex (DLPFC), frontal eye-fields and
in the NF1 sample is much higher than that, but, together with parietal cortex. Tasks performance was predicted by the activation
the presence of the genetic disorder, this could be related to in the right DLPFC.
the use of highly specific and sensitive neurocognitive diagnostic Altered inhibition relative to excitation has been reported in
techniques. other animal models of neurodevelopmental disorders in which
A previous study (Reigosa-Crespo et al., 2012) using the corticostriatal hypoactivation and working memory deficits are
same neurocognitive tests employed here, reported 9.35% of the also found. Working memory deficits have been described in
general population (children from 2 to 9th grades) exhibited Fragile X (Kwon et al., 2001), and Turner syndrome (Haberecht
Arithmetic Dysfluency (AD). The results presented here are also et al., 2001); disorders in which DD has been also detected
higher, but this could be related to the fact there are additional as part of the neurocognive phenotype. The working memory
(general domain) cognitive deficits associated to NF1 syndrome deficits in these syndromes were associated with corticostriatal
that can potencially affect arithmetical processes. Note DD chil- hypoactivation as well.
dren with NF1 exhibited spared core numeric capacities (arith- The present study didn’t include domain general cognitive
metic dysfluency was the only numeric processing impairment function evaluations, other that intellectual capacity, in order to
detected) which supports the general domain deficits hypothesis. comply with SLD diagnosis requirements. Future studies in chil-
A variety of numerical and non-numerical cognitive disabili- dren with NF1 in which SLD are detected should consider these
ties described in the general population have been described to processes, and include working memory function assessment in
have an impact in arithmetic performance: inadequate counting- order to detail the nature of the cognitive profiles described. The
based and retrieval-based strategies from long-term memory design of therapies specific to the biochemical and neurochemi-
(Geary, 1993; Jordan and Montani, 1997; Jordan et al., 2003) as cal abnormalities unique to each child would benefit significantly
well as low IQ score (Geary et al., 2000). Other domain general from this practice.
processes, generally described as deficient in NF1, such as execu- Concerning the DL findings, to the best of our knowledge,
tive functions (Bull and Scerif, 2001) and visuo-spatial working only one study examined lexical and phonological reading skills
memory (Wilson and Swanson, 2001) have been also reported to in children with NF1 (Watt et al., 2008). The authors reported
correlate with mathematical abilities. 67% of the sample exhibited deficits in one or more reading
Regarding the neural foundations of the relevant cognitive skills. Seventy-five percent of this subgroup exhibited phono-
processes, the IPS, a key structure underlying number processing logical dyslexia. Their findings indicate that a large proportion
is also believed to be involved in visuo-spatial working mem- of children with NF1 may be characterized by a specific dif-
ory, and visuo-spatial attention (Rotzer et al., 2009). Accordingly, ficulty when employing spelling-to-sound rules to assemble a
Rotzer et al. (2009) reported DD children exhibit lower behavioral pronunciation sequence.
performances in non-numerical visuo-spatial working memory The present study didn’t detect pure phonological deficits in
abilities, along with lower activity in the IPS. The authors sug- the sample analyzed. This could be related to the fact that Spanish
gest these poor spatial working memory processes may inhibit the is a transparent language (Seymour et al., 2003). On the other
formation of spatial number representations and the storage and hand, the prevalence of this subtype in the general population in
retrieval of arithmetical facts. The recruitment of extra-parietal Cuba is also significantly smaller than the corresponding to the
regions (found upon contrasting children vs. adults and children dyslexical subtype (Reigosa et al., 2008). Considering the sam-
with and without DD) also supports the contribution of gen- ple size of this study, it is statistically improbable to detect cases
eral abilities such as working memory, in numeric and arithmetic exhibiting this subtype.
achievement (Kaufmann et al., 2011). The double dissociation found here in NF1 children between
In addition, a recent translational study revealed working DD and DL supports the existence of different modular cogni-
memory deficits in Nf1 mice and in NF1 patients and hypoac- tive and neural systems, specialized in reading and numerical
tivation of related cortical striatal networks (Shilyansky et al., processing. From the neuropsychological perspective, evidence of
2010). The studies with NF1 patients showed that the degree of double dissociations between processing deficits would strongly
hypoactivation of corticostriatal networks in NF1 was predic- support the relative autonomy of the implicated processes and
tive of the degree of impairment in working memory tasks. The therefore, of the neural networks underlying them. This is espe-
convergent cross-species findings suggest hypoactivation of cor- cially relevant considering that NF1 is a genetic condition, which
ticostriatal structures, observed in NF1 patients, may be caused suggests that the development of the circuits subserving these
by increased GABA(A) receptor signaling, an effect previously domain-specific processes is highly controlled by genetic mech-
reported in fMRI studies of healthy subjects given GABA(A) anisms. These findings add to current neurocognitive modeling
agonists (Menzies et al., 2007). Additionally, Shilyansky et al. of learning and cognitive function in typical neurodevelopment.
(2010) examined the parallel effects of Nf1 mutations across Additionally, the analysis revealed that efficiencies in men-
species in homologous neuronal circuits, using behavioral tasks tal arithmetic and lexical/phonological strategies were significant
that specifically address working memory function. They pro- predictors of individual differences in math and reading attain-
vided evidence that neurofibromin/Ras signaling in both mice ment, respectively. These findings validate the diagnosis criteria
and humans regulates working memory by modulating inhibi- used in this study, and also support the current theories of cogni-
tion in prefrontal cortical and striatal networks. NF1 individuals tive development that postulate the role of a set of basic capacities
showed significantly reduced neural activity relative to controls in on higher level, school-like skills.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 6

Orraca-Castillo et al. Learning disabilities in NF1

The gender ratios related to the presence of SLD are in accor- in NF1 and the development of intervention methods for the
dance with reports in NF1 suggesting boys are at significant risk disease.
for SLD (Hyman et al., 2006). Some neurobiological evidence
could explain this finding. There have been described significantly CONCLUSIONS
less leftward asymmetry in the left planum temporale in boys The presence of SLD in the NF1 sample studied is similar to
with NF1, associated with poorer reading and math achievement. previous reports in this genetic syndrome. Interestingly, different
Significantly smaller surface area and gray matter volume were patterns of dissociation between the deficits in basic arithmeti-
also found in this area in boys, compared with girls with NF-1 cal and reading skills were detected. The present study confirmed
and controls (Billingsley et al., 2002). word decoding deficits and poor number facts retrieval under-
The gender ratios related to the presence of DD diagnosed lying DL and DD, respectively. These cognitive processes were
based on arithmetic dysfluency are much higher (5:1) than the significant predictors of reading and arithmetical attainment.
corresponding ratios in the general population reported in the However, deficits in core numeric capacities were not found in
prevalence study conducted by Reigosa-Crespo et al. (2012), the children with DD. This finding suggests that in NF1, deficits
where no difference was found between boys and girls. On the in basic number skills are not responsible for arithmetic dysflu-
other hand, the ratios corresponding to DD diagnosed on the ency. Rather, this could be explained by general-domain cognitive
basis of core numeric deficits were similarly high (4:1) to the deficits including defective working memory.
NF1 sample. This contrast in gender preponderance suggests The availability of specific neurocognitive profiles in NF1
arithmetical dysfluence in NF1 doesn’t depend on environmen- would increase the efficiency of screening methods for SLD in
tal factors, such as lack of exposure of arithmetical facts or poor genetic populations. Additionally, it would inform the design
schooling (Geary et al., 2000), as could be the case in the (typi- of individual-based cognitive interventions and pharmacological
cal) general school-age population. Rather, the cognitive deficits therapies for these children.
underlying arithmetic dysfluency in NF1 seem to be biologically
determined, as in the case of DD associated to pure core numeric AUTHOR CONTRIBUTIONS
deficits. Conceived and designed the study: Miladys Orraca-Castillo,
The M:F ratio found for DL suggests boys with NF1 are as Nancy Estévez-Pérez, Vivian Reigosa-Crespo. Performed the
likely as girls to exhibit DL. In contrast, the gender ratio in sam- experiments: Miladys Orraca-Castillo. Analyzed the data: Miladys
ples of children with reading difficulties range from 2:1 to 15:1 Orraca-Castillo, Nancy Estévez-Pérez. Wrote the manuscript:
males to females. It has been suggested that a greater variance Nancy Estévez-Pérez, Miladys Orraca-Castillo, Vivian Reigosa-
of reading performance measures in males may account at least Crespo.
in part for their higher prevalence of reading difficulties as well
as for the higher gender ratios reported. However, in research-
REFERENCES
identified samples, gender ratios are closer to 1:1 (Hawke et al., Acosta, M. T. (2007). The neurobiology of learning difficulties: neurofibromatosis
2009). type 1 as a model for researching and treating learning disorders. Rev. Neurol.
The prevalence of generalized learning disabilities associated 44(Suppl. 2), S3–S8.
with low intellectual capacity in the sample is much higher than Acosta, M. T., Bearden, C. E., Castellanos, X. F., Cutting, L., Elgersma, Y., Gioia,
G., et al. (2012). The learning disabilities network (LeaDNet): using neurofi-
that reported in reviewed publications. Between 3 and 8% of peo- bromatosis type 1 (NF1) as a paradigm for translational research. Am. J. Med.
ple with NF1 show intellectual disability (Ferner et al., 2007; Pride Genet. A 158A, 2225–2232. doi: 10.1002/ajmg.a.35535
et al., 2010; Acosta et al., 2012). Nevertheless, as in the case of Badian, N. A. (1983). “Arithmetic and nonverbal learning,” in Progress in Learning
arithmetic dysfluency deficits found in this sample, this could be Disabilities, Vol. 5, ed H. R. Myklebust (New York, NY: Grune and Stratton),
235–264.
associated with other cognitive domain general processes affect-
Billingsley, R. L., Jackson, E. F., Slopis, J. M., Swank, P. R., Mahankali,
ing intellectual capacity. For instance, there has been reported S., and Moore, B. D. (2004). Functional MRI of visual-spatial pro-
working memory explains 54% of the variance in the Raven’s cessing in neurofibromatosis, type I. Neuropsychologia 42, 395–404. doi:
Colored Progressive Matrices Test (De Ribaupierre and Lecerf, 10.1016/j.neuropsychologia.2003.07.008
2006). Additionally, in accordance with previous studies, despite Billingsley, R. L., Schrimsher, G. W., Jackson, E. F., Slopis, J. M., and Moore,
B. D. (2002). Significance of planum temporale and planum parietale mor-
neurocognitive impairments are among the most frequent symp-
phologic features in neurofibromatosis type 1. Arch. Neurol. 59, 616–622. doi:
toms of the disease, some NF1 children did not show cognitive 10.1001/archneur.59.4.616
deficits (North, 2000; Ferner et al., 2007). Bruandet, M., Molko, N., Cohen, L., and Dehaene, S. (2004). A cognitive character-
As a final consideration, note the sample size included in this ization of dyscalculia in Turner syndrome. Neuropsychologia 42, 288–298. doi:
study is relatively small. This was determined by the very low 10.1016/j.neuropsychologia.2003.08.007
Bull, R., and Scerif, G. (2001). Executive functioning as a predictor of chil-
prevalence of the NF1, thus, the results and implications of this dren’s mathematics ability: inhibition, switching, and working memory. Dev.
study must be taken with some caution. Additionally, distinct Neuropsychol. 19, 273–293. doi: 10.1207/S15326942DN1903_3
NF1-associated clinical manifestations characterized the sample Butterworth, B. (2010). Foundational numerical capacities and the origins of
and no data concerning the tumor suppressor gene inactivation dyscalculia. Trends Cogn. Sci. 14, 534–541. doi: 10.1016/j.tics.2010.09.007
timing or in what cell type’s inactivation of the NF1 gene occurs is Butterworth, B., and Reigosa-Crespo, V. (2007). “Information processing deficits
in dyscalculia,” in Why is Math So Hard for Some Children? The Nature
available. This information would allow selecting more homoge- and Origins of Mathematical Learning Difficulties and Disabilities, eds D.
nous samples, and would critically influence the understanding B. Berch and M. M. M. Mazzocco (Baltimore, MD: Paul H. Brookes),
of the biological context underlying the cognitive architecture 65–81.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 7

Orraca-Castillo et al. Learning disabilities in NF1

Caramazza, A. (1986). On drawing inferences about the structure of normal cogni- Mazzocco, M. M. (2001). Math learning disability and math LD subtypes: evidence
tive systems from the analysis of patterns of impaired performance: the case from studies of Turner syndrome, fragile X syndrome, and neurofibromatosis
for single-patient studies. Brain Cogn. 5, 41–66. doi: 10.1016/0278-2626(86) type 1. J. Learn. Disabil. 34, 520–533. doi: 10.1177/002221940103400605
90061-8 Menzies, L., Ooi, C., Kamath, S., Suckling, J., McKenna, P., Fletcher, P., et al. (2007).
Clements-Stephensa, A. M., Rimrodt, S. L., Gaur, P., and Cutting, L. E. Effects of gamma-aminobutyric acid-modulating drugs on working memory
(2008). Visuospatial processing in children with neurofibromatosis type 1. and brain function in patients with schizophrenia. Arch. Gen. Psychiatry 64,
Neuropsychologia 46, 690–697. doi: 10.1016/j.neuropsychologia.200709.013 156–167. doi: 10.1001/archpsyc.64.2.156
Crocker, L., and Algina, J. (1986). Introduction to Classical and Modern Test Theory. Molko, N., Cachia, A., Riviere, D., Mangin, J. F., Bruandet, M., Lebihan, D., et al.
New York, NY: Harcourt Brace Jovanovich College Publishers. (2004). Brain anatomy in turner syndrome: evidence for impaired social and
De Ribaupierre, A., and Lecerf, T. (2006). Relationships between working memory spatial-numerical networks. Cereb. Cortex 14, 840–850. doi: 10.1093/cercor/
and intelligence from a developmental perspective: convergent evidence from a bhh042
neo-piagetian and a psychometric approach. Eur. J. Cogn. Psychol. 18, 109–137. Murphy, M. M., and Mazzocco, M. M. (2008). Mathematics learning disabilities in
doi: 10.1080/09541440500216127 girls with fragile X or Turner syndrome during late elementary school. J. Learn.
Desoete, A., Roeyers, H., and De Clercq, A. (2004). Children with mathe- Disabil. 41, 29–46. doi: 10.1177/0022219407311038
matics learning disabilities in Belgium. J. Learn. Disabil. 37, 50–61. doi: North, K. (2000). Neurofibromatosis type 1. Am. J. Med. Genet. 97, 119–127. doi:
10.1177/00222194040370010601 10.1002/1096-8628(200022)97:2%3C119::AID-AJMG3%3E3.0.CO;2-3
Diggs-Andrews, K. A., and Gutmann, D. H. (2013). Modeling cognitive North, K. N., Riccardi, V., Samango-Sprouse, C., Ferner, R., Moore, B., Legius,
dysfunction in neurofibromatosis-1. Trends Neurosci. 36, 237–247. doi: E., et al. (1997). Cognitive function and academic performance in neurofibro-
10.1016/j.tins.2012.12.002 matosis. 1: consensus statement from the NF1 cognitive disorders task force.
Feigenson, L., Dehaene, S., and Spelke, E. (2004). Core systems of number. Trends Neurology 48, 1121–1127. doi: 10.1212/WNL.48.4.1121
Cogn. Sci. 8, 307–314. doi: 10.1016/j.tics.2004.05.002 Payne, J. M., Moharir, M. D., Webster, R., and North, K. N. (2010). Brain
Ferner, R. E., Huson, S. M., Thomas, N., Moss, C., Willshaw, H., Evans, D. G., structure and function in neurofibromatosis type 1: current concepts
et al. (2007). Guidelines for the diagnosis and management of individuals with and future directions. J. Neurol. Neurosurg. Psychiatry 81, 304–309. doi:
neurofibromatosis 1. J. Med. Genet. 44, 81–88. doi: 10.1136/jmg.2006.045906 10.1136/jnnp.2009.179630
Geary, D. C. (1993). Mathematical disabilities: cognitive, neuropsychological, Pedhazur, E., and Pedhazur-Schmelkin, L. (1991). Measurement, Design and
and genetic components. Psychol. Bull. 114, 345–362. doi: 10.1037/0033- Analysis: An Integrated Approach. Hillsdale, NJ: Lawrence Erlbaum Associates.
2909.114.2.345 Pride, N., Payne, J. M., Webster, R., Shores, E. A., Rae, C., and North, K.
Geary, D. C., Hamsom, C. O., and Hoard, M. K. (2000). Numerical and arith- N. (2010). Corpus callosum morphology and its relationship to cognitive
metical cognition: a longitudinal study of process and concept deficits in function in neurofibromatosis type 1. J. Child Neurol. 25, 834–841. doi:
children with learning disability. J. Exp. Child Psychol. 77, 236–263. doi: 10.1177/0883073809350723
10.1006/jecp.2000.2561 American Psychological Association. (1994). Diagnostic and Statistical Manual
Gross-Tsur, V., Manor, O., and Shalev, R. S. (1996). Developmental dyscalculia: of Mental Disorders, 4th Edn. Washington, DC: American Psychological
prevalence and demographic features. Dev. Med. Child Neurol. 38, 25–33. doi: Association.
10.1111/j.1469-8749.1996.tb15029.x Rasmussen, S. A., and Friedman, J. M. (2013). NF1 Gene and Neurofibromatosis 1.
Haberecht, M. F., Menon, V., Warsofsky, I. S., White, C. D., Dyer-Friedman, Am. J. Epidemiol. 151, 33–40. doi: 10.1093/oxfordjournals.aje.a010118
J., Glover, G. H., et al. (2001). Functional neuroanatomy of visuo-spatial Raven, J. C., Court, J., and Raven, J. (1992). Standard Progressive Matrices Oxford:
working memory in turner syndrome. Hum. Brain Mapp. 14, 96–107. doi: O.P. Press.
10.1002/hbm.1044 Reigosa-Crespo, V., Pérez-Abalo, M. C., Manzano, M., and Antelo, J. M. (1994).
Hawke, J. L., Olson, R. K., Willcut, E. G., Wadsworth, S. J., and Defries, J. Sistema automatizado para explorar la lectura en escolares de habla hispana.
C. (2009). Gender ratios for reading difficulties. Dyslexia 15, 239–242. doi: Revista Latina de Pensamiento y Lenguaje 2, 134–141.
10.1002/dys.389 Reigosa-Crespo, V., Valdes-Sosa, M., Butterworth, B., Estevez, N., Rodriguez,
Hyman, S. L., Arthur, S. E., and North, K. N. (2006). Learning disabilities M., Santos, E., et al. (2012). Basic numerical capacities and prevalence of
in children with neurofibromatosis type 1: subtypes, cognitive profile, and developmental dyscalculia: the havana survey. Dev. Psychol. 48, 123–135. doi:
attention-deficit-hyperactivity disorder. Dev. Med. Child Neurol. 48, 973–977. 10.1037/a0025356
doi: 10.1017/S0012162206002131 Reigosa, V., Perez Abalos, M. C., Piñeiro, A., and y López, I. (2002). Fiabilidad y
Hyman, S. L., Shores, A., and North, K. N. (2005). The nature and frequency validez de una batería de pruebas computarizadas (BTL) para la evaluación cog-
of cognitive deficits in children with neurofibromatosis type 1. Neurology 65, nitiva de la lectura. Revista Latina de Pensamiento y Lenguaje y Neuropsychologia
1037–1044. doi: 10.1212/01.wnl.0000179303.72345.ce Latina 10, 229–248.
Jordan, N. C., Hanich, L. B., and Kaplan, D. (2003). A longitudinal study of mathe- Reigosa, V., Valdes Sosa, M., Butterworth, B., Torres, P., Santos, E., Suarez,
matical competencies in children with specific mathematics difficulties versus R., et al. (2008). Large-scale prevalence studies of learning disabilities
children with comorbid mathematics and reading difficulties. Child Dev. 74, in Cuban school-children population. Clin. Neurophysiol. 119, e111. doi:
834–850. doi: 10.1111/1467-8624.00571 10.1016/j.clinph.2008.04.064
Jordan, N. C., and Montani, T. O. (1997). Cognitive arithmetic and problem solv- Restori, A. F., Katz, G. S., and Lee, H. B. (2009). A Critique of the IQ / achievement
ing: a comparison of children with specific and general mathematics difficulties. discrepancy model for identifying specific learning disabilities. Dev. Psychol. 4,
J. Learn. Disabil. 30, 624–634, 684. doi: 10.1177/002221949703000606 128–145. doi: 10.5964/ejop.v5i4.244
Kaufmann, L., Wood, G., Rubinsten, O., and Henik, A. (2011). Meta-analyses Richlan, F. (2012). Developmental dyslexia: dysfunction of a left hemisphere
of developmental fMRI studies investigating typical and atypical trajectories reading network. Front. Hum. Neurosci. 6:120. doi: 10.3389/fnhum.2012.
of number processing and calculation. Dev. Neuropsychol. 36, 763–787. doi: 00120
10.1080/87565641.2010.549884 Rotzer, S., Loenneker, T., Kucian, K., Martin, E., Klaver, P., and Von, A.
Kosc, L. (1974). Developmental dyscalculia. J. Learn. Disabil. 7, 164–177. doi: M. (2009). Dysfunctional neural network of spatial working memory con-
10.1177/002221947400700309 tributes to developmental dyscalculia. Neuropsychologia 47, 2859–2865. doi:
Kwon, H., Menon, V., Eliez, S., Warsofsky, I. S., White, C. D., Dyer-Friedman, J., 10.1016/j.neuropsychologia.2009.06.009
et al. (2001). Functional neuroanatomy of visuospatial working memory in frag- Seymour, P., Aro, M., and Erskine, J. M. (2003). Foundation literacy
ile X syndrome: relation to behavioral and molecular measures. Am. J. Psychiatry acquisition in European orthographies. Br. J. Psychol. 94, 143–174. doi:
158, 1040–1051. doi: 10.1176/appi.ajp.158.7.1040 10.1348/000712603321661859
Lewis, C., Hitch, G., and Walker, P. (1994). The prevalence of specific arith- Shallice, T. (1988). From Neuropsychology to Mental Structure. Cambridge:
metic difficulties and specific reading difficulties in 9- and 10-year old Cambridge University Press. doi: 10.1017/CBO9780511526817
boys and girls. J. Child Psychol. Psychiatry 35, 283–292. doi: 10.1111/j.1469- Shilyansky, C., Karlsgodt, K. H., Cummings, D. M., Sidiropoulou, K., Hardt, M.,
7610.1994.tb01162.x James, A. S., et al. (2010). Neurofibromin regulates corticostriatal inhibitory

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 8

Orraca-Castillo et al. Learning disabilities in NF1

networks during working memory performance. Proc. Natl. Acad. Sci. U.S.A. Conflict of Interest Statement: The authors declare that the research was con-
107, 13141–13146. doi: 10.1073/pnas.1004829107 ducted in the absence of any commercial or financial relationships that could be
Silani, G., Frith, U., Demonet, J. F., Fazio, F., Perani, D., Price, C., et al. (2005). construed as a potential conflict of interest.
Brain abnormalities underlying altered activation in dyslexia: a voxel based
morphometry study. Brain 128, 2453–2461. doi: 10.1093/brain/awh579 Received: 10 December 2013; accepted: 15 May 2014; published online: 06 June 2014.
Spelke, E. S., and Kinzler, K. D. (2007). Core knowledge. Dev. Sci. 10, 89–96. doi: Citation: Orraca-Castillo M, Estévez-Pérez N and Reigosa-Crespo V (2014)
10.1111/j.1467-7687.2007.00569.x Neurocognitive profiles of learning disabled children with neurofibromatosis type 1.
Temple, E. (2002). Brain mechanisms in normal and dyslexic readers. Curr. Opin. Front. Hum. Neurosci. 8:386. doi: 10.3389/fnhum.2014.00386
Neurobiol. 12, 178–183. doi: 10.1016/S0959-4388(02)00303-3 This article was submitted to the journal Frontiers in Human Neuroscience.
Watt, S. E., Shores, E. A., and North, K. N. (2008). An examination of lexical Copyright © 2014 Orraca-Castillo, Estévez-Pérez and Reigosa-Crespo. This is an
and sublexical reading skills in children with neurofibromatosis type 1. Child open-access article distributed under the terms of the Creative Commons Attribution
Neuropsychol. 14, 401–418. doi: 10.1080/09297040701595505 License (CC BY). The use, distribution or reproduction in other forums is permitted,
Wilson, K. M., and Swanson, H. L. (2001). Are mathematics disabilities due to a provided the original author(s) or licensor are credited and that the original publica-
domain-general or a domain-specific working memory deficit? J. Learn. Disabil. tion in this journal is cited, in accordance with accepted academic practice. No use,
34, 237–248. doi: 10.1177/002221940103400304 distribution or reproduction is permitted which does not comply with these terms.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 386 | 9