Neuropsychologia 65 (2014) 5662

Contents lists available at ScienceDirect

Neuropsychologia
journal homepage: www.elsevier.com/locate/neuropsychologia

Research Report

Weak language lateralization affects both verbal and spatial

skills: An fMRI study in 297 subjects
E Mellet a,b,c,n, L Zago a,b,c, G Jobard a,b,c, F Crivello a,b,c, L Petit a,b,c, M Joliot a,b,c,
B Mazoyer a,b,c, N Tzourio-Mazoyer a,b,c
a

Universit de Bordeaux, GIN, UMR 5296, F-33000 Bordeaux, France

CNRS, GIN, UMR 5296, F-33000 Bordeaux, France
c
CEA, GIN, UMR 5296, F-33000 Bordeaux, France
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 17 June 2014
Received in revised form
8 October 2014
Accepted 10 October 2014
Available online 22 October 2014

The present study reappraised the relationship between hemispheric specialization strength and
cognitive skills in a sample of 297 individuals including 153 left-handers. It additionally assessed the
interaction with manual laterality factors, such as handedness, asymmetry of hand motor skills, and
familial sinistrality. A Hemispheric Functional Lateralization Index (HFLI) for language was derived from
fMRI. Through mixture Gaussian modeling, three types of language hemispheric lateralization were
dened: typical (left hemisphere dominance with clear positive HFLI), ambilateral (no dominant
hemisphere with HFLI values close to 0), and strongly-atypical (right-hemisphere dominance with clear
negative HFLI values). Three cognitive scores were derived from 12 tests covering various aspects of
verbal and spatial cognition. Compared to both typical and strongly-atypical participants, those
ambilateral for language production had lower performances in verbal and non-verbal domains,
indicating that hemispheric specialization and cognitive skills are related in adults. Furthermore, this
relationship was independent from handedness and asymmetry for motor skills, as no interaction was
observed between these factors. On the other hand, the relationship between familial sinistrality and
cognitive skills tended to differ according to language lateralization type. In contrast to previous reports
in children, in the present adult population, we found no linear correlation between HFLI and cognitive
skills, regardless of lateralization type.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Hemispheric specialization
Cognitive skills
Handedness
Familial sinistrality

1. Introduction
Hemispheric specialization refers to a hemisphere-dependent
relationship between a cognitive function and a set of brain
structures (Herv et al., 2013). Although this feature is not uniquely
human, it is often considered a result of strong selection pressure
and is thus seen as an evolutionary advantage strongly related to the
apparition of language, which is emblematic of hemispheric specialization in humans (Bishop, 2013). However, the nature and even the
existence of this advantage continue to be debated. The manner
used to assess hemispheric specialization crucially affects the outcome and its interpretations. For example, the impacts on cognitive
abilities of handedness, manual preference and strength, and
asymmetry of manual skills have been considered to reect variations of brain lateralization for language (Annett, 2002; Leask and

n
Corresponding author at: Universit de Bordeaux, GIN, UMR 5296, F-33000
Bordeaux, France. Tel.: 33 5 47 30 44 01; fax: 33 5 47 30 43 94.
E-mail address: emmanuel.mellet@u-bordeaux.fr (E. Mellet).

http://dx.doi.org/10.1016/j.neuropsychologia.2014.10.010
0028-3932/& 2014 Elsevier Ltd. All rights reserved.

Crow, 2006). In this framework, Crow et al. (1998) suggested an

association between equal skill with the right and left hand and
lower abilities in both the verbal and mathematical domains,
supporting an advantage of hemispheric differentiation. However,
it is now established that relationships between handedness and
hemispheric specialization for language are far from univocal
(Knecht et al., 2000; Mazoyer et al., 2014; Szaarski et al., 2012).
Using divided visual eld presentation (Hellige, 1990), a series of
behavioral studies in healthy humans have addressed the potential
advantage of hemispheric lateralization in the language and visuospatial domains (Boles et al., 2008; Chiarello et al., 2009; Hirnstein et
al., 2010). In these paradigms, a difference in performance following
stimuli presentation in the left hemield (right hemisphere) or right
hemield (left hemisphere) is interpreted as an index reecting the
hemisphere's superiority for a given process. These studies reported
divergent outcomes relating to the language domain. Some emphasized a positive correlation between reading skills and index of
lateralization word recognition tasks, while correlations were not
signicant for the semantic tasks (Chiarello et al., 2009). In contrast,
others showed that high degrees of lateralization are detrimental to

E Mellet et al. / Neuropsychologia 65 (2014) 5662

cognitive performance in a word-matching and face-decision task

(Hirnstein et al., 2010).
While, to some extent, the divided visual eld techniques can
reportedly reveal the dominant hemisphere for language production
(Van der et al., 2011), it remains an indirect estimation of hemispheric
lateralization for language. Some studies have directly assessed hemispheric dominance using functional transcranial Doppler sonography
(fTCD) or functional magnetic resonance imaging (fMRI). However,
these methods have also produced contradictory results regarding the
relationship with performance. One study in adults derived a functional asymmetry index using fTCD during an orthographic uency
task, and found no relationship between hemispheric asymmetry and
either verbal uency, number of uently spoken foreign languages,
academic achievements, or practice of artistic activities (Knecht et al.,
2001). The authors additionally assessed general IQ in a sub-group of
21 participants, and found no relationship between functional hemispheric asymmetry and IQ. However, such an association has been
reported in developmental studies. Studies in children have demonstrated that stronger left lateralization during language production,
semantic, or phonological tasks is accompanied by better performance
in the language domain (Everts et al., 2009; Groen et al., 2012).
Among right-handed children, verbal IQ was found to be positively
associated with a laterality index in the right superior temporal gyrus
indicating stronger right-hemispheric than left-hemispheric activation in individuals with higher vs. lower verbal IQ (Lidzba et al., 2011).
Based on these ndings, one might speculate that functional
asymmetry is associated with cognitive skills among children but that
this relationship is no longer present in adults. Critically, previous
studies have focused on relationships between hemispheric lateralization and performance, without also investigating the interaction of
manual laterality factors, such as handedness, manual preference
strength, or familial sinistrality. In fact, prior studies have included
few or no left handers. We recently showed that manual laterality
factors impact performance in verbal and spatial domains through
complex interactions (Mellet et al., 2014). It has also been suggested
that hemispheric lateralization is related to manual preference
strength and familial sinistrality (Tzourio-Mazoyer et al., 2010).
Thus, it seems important to investigate these manual laterality factors
and whether they interact with hemispheric lateralization or are
independently associated with performances.
Our brief review of behavioral and neuroimaging studies highlights
that the association between cognitive skills and degree of language
hemispheric lateralization remains unclear in adults, and that its
relationship with manual laterality factors is unknown. Here we aimed
to address these issues by using the BIL&GIN (Petit et al., 2012),
a database especially designed for the study of hemispheric lateralization, which includes a sample 297 healthy individuals, balanced for sex
and handedness. The BIL&GIN participants were repeatedly scanned
with fMRI while executing an extensive battery of cognitive tasks,
including language production. The goal of this study was to establish
whether there was a relationship between hemispheric lateralization
for language and various aspects of verbal and spatial domains. We
also investigated whether the interaction between manual laterality
factors and language hemispheric lateralization impacted cognitive
performance. A lack of effect would mean that these two factors were
independently associated with cognitive performance.

2. Materials and methods

2.1. Participants
Participants were recruited within the framework of the
BIL&GIN project, a multimodal imaging/psychometric/genetic
database specically designed for studying the structural and

57

functional neural correlates of brain lateralization (Petit et al.,

2012). A total of 297 adult healthy participants (145 women)
provided informed written consent to participate in the study.
The sample included 144 self-reported right-handers (RH) and
153 left-handers (LH). The mean score to the Edimburgh
Handedness Inventory was 93.5 (SD, 11.5) for RH and  63.2
(SD, 39.6) for LH. The mean age was 25.2 years (SD, 6.4 years)
and the mean education level was 15.6 years (SD, 2.3 years) of
schooling after primary school. All participants were free of any
neurological history. The study was approved by the BasseNormandie local Ethics Committee.

2.2. Data collection

All participants were examined in three sessions. The rst
session included anatomical MRI acquisition and the rst portions
of the laterality and cognitive assessments. The cognitive assessments were completed at the second session, which was conducted a few days after the rst. All tests were administered
individually under the supervision of a trained psychologist. At the
third session, fMRI was used to acquire functional imaging data.

2.3. Determination of the participant's language lateralization type

using fMRI
2.3.1. Imaging session of language production
We evaluated hemispheric dominance for language production
using an index of asymmetry derived from t-maps obtained by
contrasting a sentence generation task to the production of a list of
words. These two conditions are fully described elsewhere
(Mazoyer et al., 2014). During 1 s, subjects were presented
white-line drawings on a black screen, which were either cartoons
depicting a scene or randomly distributed pieces of white lines.
Immediately after picture presentation, if the subject saw a
cartoon, they were instructed to covertly generate a sentence
(SENT) related to the depicted scene and to press the pad when
the sentence was completed. This enabled recording of the time
that each participant took to generate each sentence. On the other
hand, if the participant saw scrambled lines, they were instructed
to covertly generate a list (LIST) of the months of the year and to
press the pad at the end. Immediately after the session, participants were asked to recall each sentence that they covertly
generated during the fMRI session, with the support of the
pictures they saw. This procedure enabled evaluation of the
average number of words of covertly generated sentences for each
participant.
2.3.2. Image acquisition and analysis
Imaging was performed on a Philips Achieva 3 Tesla MRI
scanner. The structural MRI protocol comprised a localizer scan,
a high-resolution 3D T1-weighted volume (sequence parameters:
TR 20 ms; TE 4.6 ms; ip angle 101; inversion time 800 ms;
turbo eld echo factor 65; sense factor 2; eld of
view256  256  180 mm3; isotropic voxel size 1 mm3), and
Tn2-weighted multi-slice image acquirement (Tn2-weighted fast eld
echo (Tn2-FFE), sequence parameters: TR 3500 ms; TE 35 ms; ip
angle 901; sense factor 2; 70 axial slices; isotropic voxel
size2 mm3). Functional images were acquired with wholebrain Tn2-weighted echo planar images acquisition (Tn2-EPI,
sequence parameters: 192 volumes; TR 2 s; TE 35 ms; ip
angle 801; 31 axial slices; isotropic voxel size 3.75 mm3) covering the same eld of view as the Tn2-FFE acquisition.
Image analysis was performed using SPM5 software. The T1weighted scans were normalized to a site-specic template

58

E Mellet et al. / Neuropsychologia 65 (2014) 5662

(T-80TVS) matching the MNI space, using the SPM5 segment

procedure with otherwise default parameters, which enabled
identication of gray matter, white matter, and cerebrospinal uid
for each subject. The brain volume was calculated as the sum of
these three components.
To correct the subjects' motion during fMRI runs, 192 EPI-BOLD
scans were realigned using a rigid-body registration. The EPI-BOLD
scans were then registered rigidly to the structural T2-weighted
image, which was itself registered to the T1-weighted scan. This
combination of registration matrices allowed warping of the EPIBOLD functional scans to the standard space with a trilinear
interpolation. Once in the standard space, a 6-mm FWHM Gaussian lter was applied. Finally, for each participant, we computed
the BOLD signal difference map and the associated t-map corresponding to the map SENT minus LIST contrast.
2.3.3. Determination of the Hemispheric Functional Lateralization
Index (HFLI)
For each participant, we computed a Hemispheric Functional
Lateralization Index (HFLI) for language production, using the LItoolbox applied to the SENT minus LIST individual t-map (Wilke
and Lidzba, 2007). HFLIs were computed with a bootstrap algorithm from the individual SENT minus LIST t-maps, thresholded
at t 0 (positive t-map) with a lower bootstrap sample of 5 voxels
and higher sample size of 1000 voxels, with a resample ratio of
k 0.25. The weighted mean HFLI values were reported. HFLIs
were calculated within the anatomical template mask used for
fMRI data normalization, excluding the cerebellum. The resulting
values ranged between  100 and 100, with  100 being purely
right lateralization and 100 purely left lateralization. Hemispheric rather than regional HFLI (i.e. frontal) was used because
of its better sensitivity to assess hemispheric lateralizetion for
language (Dym et al., 2011).
2.3.4. Denition of language lateralization types using Gaussian
mixture modeling of HFLI distribution for language production
This procedure has been fully described elsewhere (Mazoyer
et al., 2014). Briey, models of n Gaussian function mixture
(n ranging from 1 to 5) were t to the HFLI data distribution,
and the optimal model was selected based on the corrected Akaike
information criterion (AICc) (Akaike, 1974; Hurvich and Tsai, 1989).
The optimal model resulted in four Gaussian functions. There was
considerable overlap between the two Gaussian components with
the highest means and, thus, we pooled the latter two components, yielding a total of three types of lateralization (Fig. 1).
Individuals with a HFLI value superior to 18 were classied as
typical (N 250, 88% of RH, 78% of LH), those with an HFLI value
between  50 and 18 were considered ambilateral (N 37, 12%
of RH, 15% of LH), and those with HFLI values below  50 were
classied as strongly-atypical (N 10, no RH, 7% of LH). Table 1
presents the characteristics of the three so-dened groups.
The three groups did not differ in age (F(2,294) 1.88, p 0.16)
or educational level (F(2,294) 1.04, p 0.35). We also observed no
difference between the three groups regarding the number of
words per sentence generated during the production task (typical:
12.4 72.0, ambilateral: 12.0 71.9, strongly-atypical: 13.4 73.0;
p 0.14), or the time spent to generate a sentence (typical:
5608 7916 ms, ambilateral: 5576 71015 ms, strongly-atypical:
56767 1238 ms; p 0.95).

Fig. 1. Distribution of SENT HFLI and corresponding classication. Three groups of

individuals were distinguished based on the distribution of Hemispheric Lateralization Indices (HFLI) during sentence production minus list of words production
(SENT minus LIST): 250 strong leftward asymmetrical individuals (typical), 37 with
moderate leftward or rightward asymmetry (ambilateral) and 10 individuals with
strong rightward asymmetry (strongly-atypical).

Table 1
Characteristics of the three groups derived from Gaussian mixture modeling.

Sex (F/M)
Handedness (LH/RH)
Edinburgh score (SD)
Motor asymmetry index
FS (FS /FS-/no info)
HFLI

Typical

Ambilateral

Strongly-atypical

123/127
120/130
20.1 (82.1)
2.5 7 5.9
90/158/2
60.3 7 13.1

16/21
23/14
 9.6 (86.0)
 0.3375.4
17/20
 6.6 7 17.5

6/4
10/0
 87.4 (17.9)
 5.4 75.0
5/5
 63.7 75.4

and 153 left-handed. The groups were balanced for sex (2(1)
0.16, p 0.69, Chi-square test). This variable is referred to as
handedness in the rest of the article.
2.4.2. Hand motor performance asymmetry index
We used the nger tapping test to assess the lateralization of
hand motor performance. Each subject was instructed to hit the
button of a small counter with their left or right index nger as
many times as possible during 10 s, keeping their wrist on the
table. Each measurement was repeated thrice for each side, and
the results were averaged for each nger. An asymmetry index
was computed as follows: (RFT  LFT/(RFT LFT)n100, where RFT
and LFT represented the average scores for the right- and left-hand
nger tapping, respectively. Group mean index values were 6.3
(SD, 4.3) in RH and  2.3 (SD, 4.1) in LH. The relationship between
hand motor performance asymmetry index and self-reported
handedness was strongly signicant (t(295) 17.6; p o0.001,
Student's t test).
2.4.3. Familial sinistrality
Positive familial sinistrality (FS) was dened as the presence of
at least one LH individual among a subject's parents or siblings.
Our sample included 112 subjects with positive FS (FS ) and 183
with negative FS (FS  ). Two subjects were unable to report on
their parents' handedness.

2.4. Manual laterality variables

2.5. Assessment of cognitive skills: tests and factorial analysis

2.4.1. Self-reported handedness

Participants answered the question Are you a left-hander or a
right-hander?, with 144 declaring themselves to be right-handed

2.5.1. Tests
Participants' verbal abilities were evaluated with the following
battery of seven tests (Table 1): 1) a recall test of an 18-word list

E Mellet et al. / Neuropsychologia 65 (2014) 5662

(Rey, 1958) for verbal memory evaluation; 2) a recall test of a list of

18 pseudo-words for verbal memory evaluation with minimal
semantic associations; 3) a verb generation test for semantic
verbal uency exploration; 4) a synonym nding test for estimating vocabulary extent (Binois and Pichot, 1956); 5) a listening span
test based on spoken sentences; 6) a reading span test based on
read sentences for verbal working memory assessment (Daneman
and Carpenter, 1980; Desmette et al., 1995); and 7) a rhyming test
on visually presented pseudo-words for evaluation of graphophonemic conversion ability.
Visuospatial abilities were assessed with the following ve
tests: 1) The mental rotation test, which estimates the ability to
rotate and spatially manipulate mental images (Vandenberg and
Kuse, 1978); 2) the Corsi Block Test, which evaluates visuospatial
short-term memory abilities (Della Sala et al., 1999); 3) a homemade 3D maze test for evaluating topographic orientation skills; 4)
a canceling task for evaluating visuospatial attention and visual
exploration abilities (Weintraub and Mesulam, 1985); and 5) the
raven matrix for assessing non-verbal reasoning (Raven, 1956).
Table 2 presents the scores on the 12 tests for each group.
2.5.2. Factorial analysis
Principal component analysis (PCA; Promax rotation) was
performed to reduce the data matrix of standardized scores from
the 12 verbal and visuospatial tests. We used the Kaiser criterion
to determine the number of factors to include (i.e. eigenvalue 41).
This resulted in a set of three components that explained 52% of
the total variance. The rst was a spatial component (SPA) that
aggregated the mental rotations test, the maze test, the Corsi Block
Test, and the Raven matrix (loading factors: 0.75, 0.71, 0.42, and
0.61, respectively). Although spatial by nature, the canceling task
was only marginally represented in this component (loading
factor: 0.16). Secondly, there was a language component (LANG)
that included the vocabulary assessment, the verbal uency task,
the reading and listening span tests, and the rhyme judgment task
(loading factors: 0.58, 0.62, 0.47, 0.48, and 0.39, respectively).
Thirdly, there was a memory component (MEM), including the
recall tests of list of words and pseudo-words (loading factor: 0.50
and 1.0, respectively). The reading and the listening span tests had
small loading factors on the MEM component (0.11 and 0.08,
respectively).
2.6. Relationships between SPA, LANG, MEM, and language
lateralization type and laterality factors

59

lateralization type by FS, and language lateralization type by motor

asymmetry index. Age, education level, sex, and brain volume were
included as confounding factors. Subjects with unknown FS status
were excluded from this analysis. Statistical analyses were performed
with JMP software (SAS, Cary, USA, version 11.0).
Other investigators have previously observed signicant correlations between HFLI and cognitive performance among children
(Everts et al., 2009; Lidzba et al., 2011). Thus, we investigated the
correlation between three components and HFLI values through a
MANCOVA performed separately for each group of language
lateralization type. The interaction of HFLI with handedness was
entered as a factor in the typical and ambilateral subgroups. Each
MANCOVA included the same confounding factors as in the rst
statistical analysis.

3. Results
The main effects of laterality factors on cognitive scores in an
extended sample have been described elsewhere (Mellet et al.,
2014). Accordingly, the present report will focus on the effects
(main and interactions) of language lateralization type on cognitive scores.
3.1. Relationships between SPA, LANG, MEM, and language
lateralization type and laterality factors
We found a main effect of language lateralization type (F
(2,282) 3.96; p 0.020) that did not differ between the three
PCA components (F(2,282) 1.14; p 0.32, Fig. 2).
Fig. S1 (see Supplementary material) displays the normalized
scores at the tests level. Although language lateralization type had
only a small effect (partial 2 0.03), it was consistent for all
scores. Post-hoc analysis revealed that ambilateral individuals
performed worse than either typical (F(1,282) 6.18; p 0.013) or
strongly-atypical subjects (F(1,282) 3.88; p 0.05). Scores did not
differ between typical and strongly-atypical subjects (F(1,282)
1.53; p 0.22). We found no signicant interaction of motor
asymmetry index by language lateralization type (F(1,282) 0.49;
p 0.61), with no difference in this interaction between the PCA
components (F(2,282) 1.34; p 0.26). Our results showed an
interaction between FS and language lateralization type (F
(2,282) 2.97, p 0.05). Post-hoc analysis evealed that the effect
of FS was similar between typical and ambilateral groups (F

To evaluate the associations between cognitive skills, lateralization

type, and laterality factors, the three PCA components were entered in
a MANCOVA. The following factors and interactions were considered:
language lateralization type, motor asymmetry index, FS, language
Table 2
Scores of typical, ambilateral and strongly-atypical participants (Mean7 SD).
Typical

Ambilateral

Strongly-atypical

Reading span (max: 6)

4.17 1.1
3.6 7 1.1
4.17 1.0
Listening span
4.8 7 1.1
4.4 7 1.2
4.25 7 1.3
Vocabulary (max: 44)
28.17 3.8
27.6 7 3.8
29.5 7 3.1
Rhyme judgment (max: 80)
67.8 7 5.5
66.77 5.4
67.6 7 4.3
Verbal uency
47.6 7 9.8
46.0 7 9.5
47.7 7 8.1
List of word (max: 90)
65.6 7 7.2
64.47 9.2
63.17 9.2
List of pseudo-word (max: 90) 35.6 17 10.6
31.9 7 11.2
37.5 7 9.9
Mental rotations (max: 20)
10.9 7 4.5
10.17 4.2
11.4 7 6.3
Corsi block
5.9 7 1.0
5.6 7 1.1
6.0 7 1.2
Topographic orientation
6.3 7 2.5
5.5 7 2.5
5.0 7 3.1
Raven matrix
111.17 10.3 106.97 10.2 106.47 8.3
Canceling task (max: 60)
47.17 8.0
43.9 7 7.7
50.9 7 6.2

Fig. 2. Scores on the 3 components in the typical, ambilateral and strongly-atypical

groups. Ambilaterals exhibited lower scores than both typicals and strong-atypicals
(p 0.013 and p 0.05 respectively). Scores did not differ between typical and
strongly-atypical subjects (p 0.22). Error bars represent the standard error of the
mean. SPA: spatial component, LANG: language component, MEM: memory
component.

60

E Mellet et al. / Neuropsychologia 65 (2014) 5662

(1,282) 0.54; p 0.46), with FS  performing better than FS in

both groups. On the other hand, the effect of FS was found to be in
the opposite direction among strongly-atypical participants compared to both the typical (F(1,282) 5.13; p 0.024) and ambilateral groups (F(1,282) 5.81; p 0.017) (Fig. 3).
The effects of handedness could not be investigated in this
analysis because all strongly-atypical participants were lefthanders. Thus, to assess these effects, we conducted an additional
MANCOVA after pooling strongly-atypical and ambilateral individuals into a single group. The three components were entered as
dependent variables, and lateralization type, handedness and
language lateralization type by handedness were entered as
independent variables. This analysis did not reveal any interaction
between language lateralization type (two modalities) and handedness (F(1,291) 0.19; p 0.67).
3.2. Relationships between SPA, LANG, and MEM scores, and strength
of cerebral lateralization for language (HFLI) in each language
lateralization type group
We found no signicant associations between HFLI and components scores in the typical (F(1,242)0.80; p 0.37), ambilateral
(F(1,29) 1.28; p 0.27), or strongly-atypical group (F(1,4) 1.70;
p 0.26). Fig. 4 presents the post-hoc correlations.
We found no signicant difference between the three PCA
components in any of the language lateralization groups [typical:
F(2,241) 0.80, p 0.45; ambilateral: F(2,28) 0.66, p 0.52; and
strongly-atypical: F(2,3) 0.45, p 0.57. Similarly, there was no
interaction between HFLI and handedness on the PCA components
in either typical (F(1,242)0.14; p 0.71) or ambilateral individuals (F(1,29) 0.63; p 0.43). This absence of interaction was
similar among the three components (F(2,241) 2.34, p 0.10 in
typicals; F(2,28) 0.25, p 0.78 in ambilaterals).

4. Discussion
The present work aimed to assess the relationships between
hemispheric lateralization for language and performance in various aspects of verbal and spatial cognition.
The originality of this study lies in several features. First, the
sample included 297 healthy subjects balanced for handedness,
thus comprising a much higher proportion and number of lefthanders than previous studies. This enabled us to account for the

Fig. 3. Interaction between language lateralization type and familial sinistrality.

The effect of familial sinistrality (FS) was found to be in the opposite direction
among strongly-atypical participants compared to both typical and ambilateral
groups. The mean component score of the y-axis corresponds to the average of the
values of SPA, LANG and MEM components. Errors bars represent the standard
error of the mean.

Fig. 4. Correlation between HFLI and mean component score. No correlation was
found in any group. Typicals: r 0.05, p 0.37, ambilaterals: r 0.22, p 0.18,
strongly-atypicals: r 0.11, p 0.76. The mean component score of the y-axis
corresponds to the average of the values of SPA, LANG and MEM components.

higher variability of language lateralization type among lefthanders. Secondly, the brain lateralization index for language
was assessed using fMRI, an approach that is as accurate as the
Wada test (Binder, 2011; Dym et al., 2011). The reliability of the
lateralization index was further improved by contrasting a sentence production task with a high-level reference condition, such
as the automatic production of a list of words, (Binder, 2011).
Moreover, the classication of participants into three types of
lateralization resulted from unsupervised Gaussian mixture modeling, while other studies have based such classication on an
arbitrary threshold applied to the index distribution (Everts et al.,
2009; Groen et al., 2012; Knecht et al., 2001; Lidzba et al., 2011).
Thirdly, the present work included a large set of tests aimed at
assessing cognitive abilities in both verbal and spatial domains.
The principal components derived from these tests thus involved
most aspects of these cognitive domains. Finally, in addition to
manual preference, the present analysis included other laterality
factors known to affect cognitive performance, such as asymmetry

E Mellet et al. / Neuropsychologia 65 (2014) 5662

of motor performance and familial sinistrality (Mellet et al., 2014).

This made it possible to unravel potential effects on cognitive
performances of the interactions between these factors and the
type of hemispheric lateralization for language.
Our present results showed that subjects weakly lateralized for
language production performed worse in the verbal, verbal memory, and spatial components. Strikingly, the effect on performance
of weak or absent lateralization for language production was not
conned to the verbal domain but also equally affected spatial
cognition. Lateralization for language production can reportedly be
correlated with skills in other verbal domains (Chiarello et al.,
2012, 2009). Our results extend this knowledge, by suggesting that
the index of lateralization for language production reects variations in the global organization of the brain rather than solely in
the organization of language.
Researchers using hand skill asymmetry measurement as an
indirect marker of brain lateralization have previously proposed
the existence of a so-called hemispheric indecision point, around
which participants had lower academic abilities (Crow et al., 1998;
Peters et al., 2006), although this concept has been challenged by
others (Mayringer and Wimmer, 2002). In this regard, our results
provide the rst evidence that near-zero values of brain activation
asymmetry for language production are indeed associated with
lower performance in various cognitive domains. This nding goes
against the hypothesis that strongly lateralized individuals should
exhibit lower performances than less lateralized ones (Annett and
Manning, 1989).
It must be noted that the present outcome does not agree with
that of a previous fTCD study reporting no association between
scores and strength of hemispheric lateralization (Knecht et al.,
2001). This discrepancy may be explained by two aspects of the
previous work. First, in the study by Knecht et al. lateralization
types were derived from the standard error of the lateralization
index, leading to a larger proportion of participants categorized as
strongly-atypical (10% vs. 3% in the present study). This different
denition of language lateralization type could have blurred the
group differences. Second, most abilities assessed by Knecht et al.
are hardly comparable with the cognitive evaluation implemented
in the present work. The authors particularly focused on academic
achievements and the number of uently spoken foreign languagescharacteristics that are more closely related to education
level than to well-dened cognitive skills. In this regard, it is
notable that our present results showed similar education levels
among the typical, ambilateral, and strongly-atypical participants.
Knecht et al. additionally measured IQ and linguistic processing
speed in a sub-group of 21 subjects (7 subjects of each group), and
found that these scores were not correlated with the strength of
lateralization. However, these negative results could be due to a
lack of sensitivity. A more recent study in 36 right-handed children
did not reveal any association between verbal IQ and hemispheric
lateralization during a covert word generation task (Lidzba et al.,
2011). However, the asymmetry was assessed in only few and
relatively small regions of interest located in the inferior frontal,
precentral, and middle temporal gyri. In contrast, the present work
considered the whole hemispheric asymmetry. Interestingly, a
study that used larger regions of interest did report a positive
correlation (Everts et al., 2009).
A recent study found that hemispheric lateralization of verbal
uency and of visuospatial processing were not related to performances in verbal comprehension and perceptual organization
(Powell et al., 2012). However, the authors reported an effect of
the interaction between the two hemispheric lateralizations, with
performances in both verbal and perceptual domains being lower
when the verbal and visuospatial processes were associated in the
same hemisphere. It is possible that this hemispheric organization
concern preferentially ambilateral participants, what could partly

61

explain the presently observed relationship between hemispheric

lateralization for verbal production and performances in the
spatial domain.
Although the small number of strong-atypical subjects calls for
caution, our present results suggest that the relationship between
lateralization for language production and cognitive performances
is not directional. The strongly lateralized individuals performed
better than weakly lateralized ones, regardless of the side of
lateralization. This is in line with the conclusion of a metaanalysis of divided elds studies including various language tasks
(Boles et al., 2008). Notably, the ambilateral participants had no
history of specic language impairment. This underlines that
although decreased leftward lateralization has been observed in
dyslexic participants, ambilaterality could also be a normal variant
of language lateralization (Bishop, 2013; Illingworth and Bishop,
2009).
We found no association between HFLI and scores, regardless of
lateralization type. This is in contrast to the linear relationship
between hemispheric lateralization index and verbal and nonverbal performances that has been previously reported in developing children (Everts et al., 2009), This discrepancy may suggest
that once brain maturation is completed and lateralization has
reached a plateau (Szaarski et al., 2006), there is no longer a
correlation between performance and the degree of lateralization,
although ambilateral subjects can still be distinguished from
typical and strongly-atypical subjects. These two observations
are not contradictory. It is indeed possible that these differences
between groups reect variability in the establishment of hemispheric lateralization during brain development. Later, the relationship between hemispheric lateralization and cognitive skills
could have been altered by external factors (e.g. education) that
have a more important effect on performance than behavioral
manual laterality factors (Mellet et al., 2014).
Our present results showed no interaction between hemispheric lateralization for language and either handedness or
asymmetry in manual ability. Handedness-related literature typically confounds the effects of behavioral laterality such as
handedness and asymmetry of manual ability with variations
of hemispheric lateralization (Annett, 1992; Annett and Manning,
1989; Crow et al., 1998; Leask and Crow, 2001). However, our
ndings indicate that the association between language lateralization and cognitive skills is unrelated to behavioral manual laterality. This is consistent with the reportedly weak association
between handedness and lateralization for language (Mazoyer et
al., 2014).
Here we found that familial sinistrality was the unique laterality factor that showed a signicant interaction with the type of
language lateralization, resulting in different effects in stronglyatypical compared to either typical or ambilateral participants.
In the former group, familial sinistrality was not associated with
weaker performances as previously reported (DAndrea and Spiers,
2005; Mellet et al., 2014; Searleman et al., 1984). Obviously, this
result must be interpreted cautiously given the small number of
strongly-atypical participants. However, it emphasizes the fact
that, although handedness is weakly heritable, familial sinistrality
not only affect performances but also shapes some anatomical and
functional features in the brain (Tzourio-Mazoyer et al., 2010,
2011).
In summary, the present results suggest that the strength but not
the direction of hemispheric specialization for language relates to
cognitive skills in both verbal and spatial domains. The rationale
behind this association remains speculative as there is no hint of a
causal relationship. Better characterization of this relationship will
require a study including more strongly-atypical subjects, and thus
with a considerable number of participants, as they constitute less
than 1% of the general population. Several developmental studies have

62

E Mellet et al. / Neuropsychologia 65 (2014) 5662

demonstrated that verbal and spatial competences independently

account for an increase in laterality with age. The relative decrease
in performance among adult ambilateral subjects might constitute a
remnant of developmental variability during the establishment of
hemispheric specialization.
Acknowledgments
The authors are deeply indebted to G. Perchey for his help with
fMRI acquisition, and to I. Hessling for her careful reading and
comments.
Appendix A. Supporting information
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.neuropsychologia.
2014.10.010.
References
Akaike, H., 1974. A new look at the statistical model identication. IEEE Trans.
Autom. Control 19 (6), 716723.
Annett, M., 1992. Spatial ability in subgroups of left-and right-handers. Br. J.
Psychol. 83 (4), 493515.
Annett, M., 2002. Handedness and Brain Asymmetry: The Right Shift Theory.
Psychology Press, New York.
Annett, M., Manning, M., 1989. The disadvantages of dextrality for intelligencen. Br.
J. Psychol. 80 (2), 213226.
Binder, J.R., 2011. Functional MRI is a valid noninvasive alternative to wada testing.
Epilepsy Behav.: E&B 20 (2), 214222.
Binois, D.R., & Pichot, D.P, 1956. Test de vocabulaire: manuel d'application. Editions
du Centre de psychologie appliquee.
Bishop, D.V.M., 2013. Cerebral asymmetry and language development: cause,
correlate, or consequence? Science 340 (6138), 1230531.
Boles, D.B., Barth, J.M., Merrill, E.C., 2008. Asymmetry and performance: toward a
neurodevelopmental theory. Brain Cogn. 66 (2), 124139.
Chiarello, C., Welcome, S.E., Leonard, C.M., 2012. Individual differences in reading
skill and language lateralisation: a cluster analysis. Laterality 17 (2), 225251.
Chiarello, C., Welcome, S.E., Halderman, L.K., Leonard, C.M., 2009. Does degree of
asymmetry relate to performance? An investigation of word recognition and
reading in consistent and mixed handers. Brain Cogn. 69 (3), 521530.
Crow, T., Crow, L., Done, D., Leask, S., 1998. Relative hand skill predicts academic
ability: global decits at the point of hemispheric indecision. Neuropsychologia
36 (12), 12751282.
DAndrea, E.A., Spiers, M.V., 2005. The effect of familial sinistrality and academic
experience on cognition in right-handed women. Neuropsychology 19 (5), 657.
Daneman, M., Carpenter, P.A., 1980. Individual differences in working memory and
reading. J. Verbal Learn. Verbal Behav. 19 (4), 450466.
Della Sala, S., Gray, C., Baddeley, A., Allamano, N., Wilson, L., 1999. Pattern span: a
tool for unwelding visuospatial memory. Neuropsychologia 37 (10),
11891199.
Desmette, D., Hupet, M., Schelstraete, M.A., Van der Linden, M., 1995. Adaptation en
langue franaise du reading span test de daneman et carpenter (1980).
L'anne Psychol. 95 (3), 459482.
Dym, R.J., Burns, J., Freeman, K., Lipton, M.L., 2011. Is functional MR imaging
assessment of hemispheric language dominance as good as the wada test?:
ameta-analysis. Radiology 261 (2), 446455.
Everts, R., Lidzba, K., Wilke, M., Kiefer, C., Mordasini, M., Schroth, G., et al., 2009.
Strengthening of laterality of verbal and visuospatial functions during childhood and adolescence. Hum. Brain Mapp. 30 (2), 473483.
Groen, M.A., Whitehouse, A.J., Badcock, N.A., Bishop, D.V., 2012. Does cerebral
lateralization develop? A study using functional transcranial doppler ultrasound assessing lateralization for language production and visuospatial memory. Brain Behav. 2 (3), 256269.
Hellige, J.B., 1990. Hemispheric asymmetry. Annu. Rev. Psycho. 41 (1), 5580.

Herv, P.-Y., Zago, L., Petit, L., Mazoyer, B., Tzourio-Mazoyer, N., 2013. Revisiting
human hemispheric specialization with neuroimaging. Trends Cogn. Sci..
Hirnstein, M., Leask, S., Rose, J., Hausmann, M., 2010. Disentangling the relationship
between hemispheric asymmetry and cognitive performance. Brain Cogn. 73
(2), 119127.
Hurvich, C.M., Tsai, C.-L., 1989. Regression and time series model selection in small
samples. Biometrika 76 (2), 297307.
Illingworth, S., Bishop, D.V., 2009. Atypical cerebral lateralisation in adults with
compensated developmental dyslexia demonstrated using functional transcranial doppler ultrasound. Brain Lang. 111 (1), 6165.
Knecht, S., Drger, B., Deppe, M., Bobe, L., Lohmann, H., Flel, A., et al., 2000.
Handedness and hemispheric language dominance in healthy humans 123 (12),
25122518Brain: J. Neuro. 123 (12), 25122518.
Knecht, S., Drger, B., Flel, A., Lohmann, H., Breitenstein, C., Deppe, M., et al., 2001.
Behavioural relevance of atypical language lateralization in healthy subjects
124 (8), 16571665Brain: J. Neuro 124 (8), 16571665.
Leask, S., Crow, T., 2001. Word acquisition reects lateralization of hand skill.
Trends Cogn. Sci. 5 (12), 513516.
Leask, S.J., Crow, T.J., 2006. A single optimum degree of hemispheric specialisation
in two tasks, in two UK national birth cohorts. Brain Cogn. 62 (3), 221227.
Lidzba, K., Schwilling, E., Grodd, W., Krgeloh-Mann, I., Wilke, M., 2011. Language
comprehension vs. language production: age effects on fmri activation. Brain
Lang. 119 (1), 615.
Mayringer, H., Wimmer, H., 2002. No decits at the point of hemispheric indecision. Neuropsychologia 40 (7), 701704.
Mazoyer, B., Zago, L., Jobard, G., Crivello, F., Joliot, M., Perchey, G., Mellet, E., Petit, L.,
Tzourio-Mazoyer, N., 2014. Gaussian mixture modeling of hemispheric lateralization for language in a large sample of healthy individuals balanced for
handedness. PLoS One 9 (6), e101165.
Mellet, E., Jobard, G., Zago, L., Crivello, F., Petit, L., Joliot, M., et al., 2014. Relationships between hand laterality and verbal and spatial skills in 436 healthy adults
balanced for handedness. Laterality 19 (4), 122.
Peters, M., Reimers, S., Manning, J., 2006. Hand preference for writing and
associations with selected demographic and behavioral variables in 255,100
subjects: the BBC internet study. Brain Cogn. 62 (2), 177189.
Petit L., Crivello F., Mellet E., Jobard G., Zago L., Joliot M., Perchey G., Mazoyer B.,
Tzourio-Mazoyer N., BIL&GIN: A database for the study of hemispheric
specialization, 2012, In 18th annual meeting of the organization for human
brain mapping. Beijing, China.
Powell, J.L., Kemp, G.J., Garca-Finaa, M., 2012. Association between language and
spatial laterality and cognitive ability: an fMRI study. Neuroimage 59 (2),
18181929.
Raven, J.C., 1956. Guide to the Standard Progressive Matrices. The Psychological
Corporation, New York.
Rey, A., 1958. L'examen clinique en psychologie. Presses Universitaires De France,
Paris.
Searleman, A., Herrmann, D.J., Coventry, A.K., 1984. Cognitive abilities and lefthandedness: an interaction between familial sinistrality and strength of
handedness. Intelligence 8 (4), 295304.
Szaarski, J.P., Holland, S.K., Schmithorst, V.J., Byars, A.W., 2006. FMRI study of
language lateralization in children and adults. Hum. Brain Mapp. 27 (3),
202212.
Szaarski, J.P., Rajagopal, A., Altaye, M., Byars, A.W., Jacola, L., Schmithorst, V.J., et al.,
2012. Left-handedness and language lateralization in children. Brain Res. 1433,
8597.
Tzourio-Mazoyer, N., Petit, L., Razamandimby, A., Crivello, F., Zago, L., Jobard, G.,
et al., 2010. Left hemisphere lateralization for language in right-handers is
controlled in part by familial sinistrality, manual preference strength, and head
size. J. Neurosci. 30 (40), 1331413318.
Tzourio-Mazoyer, N., Simon, G., Crivello, F., Jobard, G., Zago, L., Perchey, G., et al.,
2011. Effect of familial sinistrality on planum temporale surface and brain tissue
asymmetries. Cereb. Cortex 20 (6), 14761485.
Vandenberg, S., Kuse, A., 1978. Mental rotations, a group test of three-dimensional
spatial visualization. Percept. Motor Skills 47, 599604.
Van der Haegen, L., Cai, Q., Seurinck, R., Brysbaert, M., 2011. Further fmri validation
of the visual half eld technique as an indicator of language laterality: a largegroup analysis. Neuropsychologia 49 (10), 28792888.
Weintraub, S., Mesulam, M.M., 1985. Mental state assessment of young and elderly
adults in behavioral neurology, Principles of Behavioral Neurology, F.A. Davis
Company ed. Mesulam MM, Philadelphia, pp. 71123.
Wilke, M., Lidzba, K., 2007. LI-tool: a new toolbox to assess lateralization in
functional MR-data. J. Neurosci. Methods 163 (1), 128136.