Lexical Production and Cognitive Control in Bilinguals

In bilinguals, any production in either of the languages leads to the
activation of both languages at the same time. The mechanism of language
activation in bilinguals is mostly believed to be nonselective, thus, there is
a parallel activation of both L1 and L2 lexical representations (Kroll, Bobb, &
Wodniecka, 2006; Marian, Spivey, & Hirsch, 2003; Sunderman, & Kroll, 2006).
Such a parallel activation of both languages at the same time increases the
attentional control in bilinguals, to control the activation of non-target
lexical representations (Christoffels, Firk, & Schiller, 2007), which is
assumed to occur at the lexical level of word production (Costa, 2005;
Finkbeiner, Gollan, & Caramazza, 2006). For a detailed linguistic description
of different stages involved in word production see Bloem and La Heij (2003);
Bloem, van den Boogaard, and La Heij (2004); Levelt, Roelofs, and Meyer (1999);
Schriefers, Meyer, and Levelt (1990).
In general, those aspects of cognitive control which are involved in selecting
and maintaining a response in the face of a conflict and in the presence of
some other alternatives are characterized by the involvement of such neural
systems as prefrontal, inferior parietal cortex, anterior cingulate cortex
(ACC) (Braver, Barch, Gray, Molfese, & Avraham, 2001; Bunge, Hazeltine,
Scanlon, Rosen, & Gabrieli, 2002), as well as basal ganglia (Middleton &
Strick, 2000). Prefrontal cortex has been recognized to bring about a
facilitating processing manner via its top down bias mechanisms when irrelevant
candidates compete with those representations which are related to a task
(Miller & Cohen, 2001). Prefrontal cortex has strong interconnections with the
parietal cortex (Petrides & Pandya, 1984). Such a circuit has been reported to
play a role when there is a need to select among some competing responses, with
left parietal cortex engaged in activating responses which are possible, and
prefrontal cortex involved in selecting a response among competing candidates
(Bunge, et al., 2002). In addition to prefrontal cortex, ACC has also been
suggested to contribute to response selection. Although prefrontal cortex
contributes to response selection when there is a conflict among competing
candidates, it is the ACC which formulates the degree of cognitive control
(Bush, Luu, & Posner, 2000). In general, the amount of ACC activation depends
on the degree of conflict in selecting a response (Carter et al., 1998); in the
process of response selection, ACC identifies the conflict among competing
cues, then the prefrontal cortex via a signal received from ACC on the
existence of a conflict, exerts more control provided by the top down
regulatory mechanisms of posterior cortex or the basal ganglia (MacDonald,
Cohen, Stenger & Carter 2000). The basal ganglia may also contribute to
language planning (Fabbro, Peru, & Skrap, 1997) as well as selecting a suitable
lexical item (Wallesch, 1985).
Regarding the neural substrates involved in bilingual lexical production, there
has been a debate over the last two decades. Earlier research reported no
neural differences underlying L1 and L2 lexical production; thus a common
neural system including bilateral network in the frontal, temporal, parietal
and occipital cortex along with the cerebellum and limbic system, was mostly
emphasized (Hernandez, Dapretto, Mazziotta, & Bookheimer, 2001; Pu et al.,
2001). However, later, literature has witnessed different neural dissociations
in L1 and L2 processing. To name a few, as a result of L2 production, one can
refer to increased activation in frontal cortex (Vingerhoets et al., 2003),
prefrontal cortex (De Bleser et al., 2003), the right insula, the anterior
cingulate gyrus, the dorsolateral prefrontal cortex, and the left fusiform
gyrus (Hernandez & Meschyan, 2006), the left inferior frontal gyrus (LIFG),
bilateral supplementary motor area (SMA), left precentral gyrus, left lingual
gyrus, left cuneus, bilateral basal ganglia including the putamen, globus
pallidus, and caudate, and bilateral cerebellum (Liua, Hub, Guoa & Peng, 2010).
Therefore, it has been argued that in L2 lexical production more neural
resources are involved, possibly to overcome L1 interference with L2
production, and to inhibit L1 activation, as L2 production is less automatic
(Abutalebi et al., 2008); in contrast, more activation in the right basal
ganglia is associated with L1 production (Liua et al., 2010).
In more recent research (Abutalebi et al., 2013; Reverberi et al, 2015) the
pre-SMA/ACC, prefrontal cortex and the left caudate, known as language control
network, have been reported to be involved in L2 lexical production. It has
been mentioned that when producing words in L2 compared with L1 production, the
pattern of activation agrees with the language control network, indicating that
L2 production requires recruitment of more control processes compared with L1
(Reverberi et al., 2015); however, another network, is assumed to be associated
with L1 word production; in fact, Reverberi et al. (2015) reported that L1 word
production activates more bilaterally the inferior parietal lobules, the
precuneus, the posterior cingulate cortex, and the right lateral prefrontal
cortex. It is considered that the activity of the regions involved in the
latter network correlate negatively with the activity of the regions involved
in L2 word production (Fornito, Harrison, Zalesky, & Simons, 2012).

Thus far, what previous literature has dealt with, addresses the neural
mechanisms involved in L1 and L2 lexical production separately (e.g. Hernandez
& Meschyan, 2006; Liua et al., 2010; Reverberi et al., 2015), irrespective of
any explanations on neural mechanisms involved in reversed language effect, and
the absence of standard language effect in bilinguals. The present research not
only addresses the gap in the related state-of-the-art research, but also
investigates the underling mechanisms involved in the aforementioned effects;
such investigations contribute to understanding the primary processes involved
in bilingual first language attrition and any inspection of the role of
cognitive control in those effects provides detailed analysis of the way
bilingual brain adopts an adaptive mechanism in producing L1 and L2 lexicons.
Therefore, the main incentives of this research revolve around shedding light
on the function of neural mechanisms which bring about the reversed language
effect, and the absence of standard language effect in bilingual lexical
production and the way cognitive control modulates such language behavior.

This experiment includes a standard structural T1-weighted MRI scan, a
Diffusion Tensor Imaging (DTI), a resting state fMRI, and a picture naming task
in three sections (pre-switch, switch, post-switch). The experimental design is
a 2 (Dutch: English) x 3 (pre-switch: switch: post-switch) factorial design,
within and between participants. In this experiment, within subject analysis
will look into the contrast in different experimental conditions and between
subject analysis concentrates on any differences that could be observed when in
pre-switch and post-switch blocks, some participants receive first English
block and then Dutch block and some other participants receive this order vice
versa. Dependent measures are participants* reaction times and brain activity
in response to different conditions of the experiment. Brain activation during
task performance will be measured using standard event-related and blocked
design functional MRI. Also, correlations in the fMRI time courses during
lexical production will be measured in order to establish the functional
connectivity network related to lexical production and cognitive control.
Participants will be instructed about the experimental procedures and will be
familiarized with the task that they will do inside the scanner. The first
section of the picture naming task, pre-switch, has two conditions (English and
Dutch) in block design with 48 task trials and 80 baseline trials in two runs,
one run in Dutch and one run in English. Each run consists of three task blocks
(with the mean of eight trials in each block) and four baseline blocks (with a
mean of ten trials in each block). Task blocks and baseline blocks alternate
each other. The duration of pre-switch will be seven minutes.
In switch section, participants will be required to name the stimuli by
switching between English and Dutch. Switch section has four conditions
(English and Dutch) x (Switch and nonswitch trials) with 96 trials, in which
the inter stimulus interval will be jittered. A cue color will hint
participants in what language the next image should be named. In switch
section, the stimuli will be the ones that are used in pre-switch section. The
duration of this section will be six minutes. Based on the behavioral study
that has been done in advance by the researcher, switch section is, in fact,
where the facilitatory function over bilinguals* second language, an inhibitory
function over their first language along with reversed language effect, are
expected to occur.
Post-switch has four conditions (English and Dutch) x (new and old stimuli)
with 96 task trials and 160 baseline trials in four runs, two runs in Dutch and
two runs in English. Each run consists of three task blocks (with the mean of
eight trials in each block) and four baseline blocks (with a mean of ten trials
in each block). Task blocks and baseline blocks alternate each other. This
section will take fourteen minutes. Compared with pre-switch section,
post-switch section has two runs more. Post-switch section will include both,
stimuli used in pre-switch and new stimuli. Based on the behavioral study that
has been done in advance by the researcher, post-switch is where reversed
language effect, and the absence of standard language effect are expected to be
observed.

Participating in an fMRI and DTI study has not been associated to any known
risks. These non-invasive techniques involve no catheterizations or
introduction of exogenous tracers. Numerous children and adults have undergone
magnetic resonance studies without apparent harmful consequences. Some people
become claustrophobic while inside the magnet and in these cases the study will
be terminated immediately at the subject's request. The only absolute
contraindications to MRI studies are the presence of intracranial or
intraocular metal, or a pacemaker. Relative contraindications include pregnancy
and claustrophobia. Participants who may be pregnant, who may have metallic
foreign bodies in the eyes or head, or who have cardiac pacemakers will be
excluded because of potential contraindications of MRI in such subjects.
Although there is no direct benefit to the participants from this proposed
research, there are greater benefits to society from the potential knowledge
gained from this study.
There is no direct benefit to the participants from this research; however,
this research by addressing the gap in the related state-of-the-art research,
provides insights on the function of neural mechanisms involved in bilingual
lexical production, and cognitive control; especially given that the importance
of the benefits gained from this research far outweighs the minimal risks
involved.
In this study, participants will be protected against any MRI procedural risks
via a thorough pre-screening process. Information obtained from the study will
be strictly confidential, except as required by law, and will be made available
to the subject and his/her physician in response to a specific request from the
subject. There will be no personal identification of subjects in scientific
communications. Data will be stored in a confidential manner both through the
use of a coding system (a code will be assigned to the data from a given
subject instead of the subject*s name) and through the security of the files
and computer systems.
All standard anatomical images (i.e. localizer images) are routinely reviewed
by a neuroradiologist. fMRI studies are not given a clinical interpretation. In
the event that a significant abnormality is detected, a recommendation to seek
further medical consultation will be made. However, it is stressed that the MRI
evaluation performed for these studies does not represent a complete clinical
MRI evaluation, and it is not being performed for clinical diagnostic purposes.