Neural Basis of Language
and Set Shifting in
Bilinguals
Cesar Avila, Gabriele Garbin, Ana Sanjuan, Cristina
Forn, Juan-Carlos Bustamante, Aina Rodríguez-
Pujadas, Mireia Hernández & Albert Costa
SUMMARY
-Some reflections about Spanish-Catalan bilingualism
- Language control in bilingualism
- Study 1: Language control in Spanish-Catalan bilinguals
- Task switching in bilinguals
- Study 2: Task switching in Spanish-Catalan bilinguals and
monolinguals
- Conclusions
BrainGlot 2009 - Barcelona
GENERAL PURPOSE
At least for some kind of early and
high proficient bilinguals, language
and cognitive switching would share
similar neural mechanisms.
BrainGlot 2009 - Barcelona
Spanish-Catalan Bilingualism: Some
reflections
Both languages are similar and easy to
learn simultaneously.
There are a lot of bilinguals, with a
diverse level of proficiency in Catalan.
Bilingualism developed because a person
lives in a regional context in which two
languages are co-official
There is often a concurrent L1 and L2
development since early infancy
Intense training in both languages
BrainGlot 2009 - Barcelona
Spanish-Catalan Bilingualism:
A good scenario for training executive functions
since infancy. For example, one child can speak
or hear L1 or L2 as a function of a cue:
Catalan: father, grandparents (father), school,
some friends, some relatives....
Spanish: mother, grandparents (mother), at
kindergarten, some other friends, some other
relatives...
Plus zapping TV, reading, etc.
Involved in continous task switching....Mixing
languages becomes effortless.
BrainGlot 2009 - Barcelona
Spanish-Catalan Bilingualism: Some
reflections
Special case of bilingualism where
switching may be trained since early
infancy This implies that the
development of some executive functions
(response selection, response inhibition,
etc.) is intensively trained in parallel to
the development of language learning.
Special case for the development of a
“passive” bilingualism A good
comprehension but low expression in L2
(See poster)
BrainGlot 2009 - Barcelona
Objectives
To study the neural basis of language
control in Spanish-Catalan early bilinguals
To study the neural basis of task switching
in Spanish-Catalan early bilinguals when
compared with monolinguals
To test if both processes share the same
neural basis.
BrainGlot 2009 - Barcelona
Language control in early
bilinguals
BrainGlot 2009 - Barcelona
NEURAL BASIS OF LANGUAGE
CONTROL
Proposed by Abulatebi and Green,
2007
LIFG (Broca’s
area) and RIFG:
task switching
and language
control
BrainGlot 2009 - Barcelona
Abutalebi (2008)
BrainGlot 2009 - Barcelona
Behavioral Studies in language control
Picture Naming Task (Costa & Santesteban, 2004)
Switching bewteen L1 and L2
L1=Spanish Red pictures
L2=Catalan Blue pictures
Non-Switch
Backward Switch
Non-Switch
Forward-Switch
Backward
Switch
BrainGlot 2009 - Barcelona
PREVIOUS BEHAVIORAL STUDIES
Asymmetrical
switching costs in low
proficient bilinguals:
-Switching is slower
than non-switching
-Naming in L1 from L2
is slower than naming
in L2 from L1.
-Specific for late
bilinguals
BrainGlot 2009 - Barcelona
PREVIOUS BEHAVIORAL
STUDIES
Symmetrical switching
costs in early, high
proficient bilinguals:
-Switching is slower
than non-switching
-Naming in L1 is
slower than naming in
L2.
-Specific for early
bilinguals
independently of
proficiency (for L3)
BrainGlot 2009 - Barcelona
Language control and lexical competition in bilinguals: an
event-related fMRI study. Neuroimage Wang et al. (2007)
Participants
Twelve late bilingual students (6 females)
mean age 19.5 (from 18 to 21)
L1: Chinese
L2: English learnt at 12.67(SD + 1.2) years old
(poor proficiency score=3 over 5)
BrainGlot 2009 - Barcelona
Wang et al. (2007)
“English”
(200ms) (2800 ms)
(2800 ms) (200ms)
BrainGlot 2009 - Barcelona
Wang et al. (2007)
Forward switching Backward switching
L1 to L2 L2 to L1
BrainGlot 2009 - Barcelona
Neural bases of asymmetric language switching
in second-language learners: An ER-fMRI study
(Wang et al., 2007)
Wang’s conclusion: the neural correlates of language
switching differ depending on the direction of the switch
and there does not seem to be a specific brain area
acting as a “language switch”.
But: 3 seconds of ITI is too long to investigate language
switching
Different types of bilingualism should be considered
BrainGlot 2009 - Barcelona
Language control and lexical competition in bilinguals: an
event-related fMRI study (Abutalebi et al., 2008)
Participants
12 late bilinguals (10 females). L1: German; L2:
French; AOA= 11.6.
fMRI tasks
Simple Naming (SNc): naming pictures in L1
Task Selection (TSc): naming pictures or
generating verbs from pictures as a function of a
cue (all in L1)
Language Selection (LSc): naming pictures in L1
or L2 as a function of a cue
BrainGlot 2009 - Barcelona
Results I (Abutalebi et al., 2008)
Abutalebi et al., 2008
Naming L1 in dual vs single task: LIFG,
SMA
BrainGlot 2009 - Barcelona
Results II (Abutalebi et al., 2008)
Abutalebi et al., 2008
Naming L1 in bilingual vs monolingual task:
LIFG, RIFG, ACC, bilateral striatum
BrainGlot 2009 - Barcelona
Results III (Abutalebi et al., 2008)
Abutalebi et al., 2008
Naming L1 in bilingual vs dual task: LIFG,
RIFG, ACC and left striatum
BrainGlot 2009 - Barcelona
Results IV (Abutalebi et al., 2008)
Naming L2 vs L1 in bilingual context: LIFG,
RIFG, ACC and bilateral striatum
BrainGlot 2009 - Barcelona
Discussion
Results have confirmed Abutalebi and
Green’s model involving LIFG, RIFG, ACC
and the striatum in language control.
Directional changes from L1 to L2 and
vice-versa, were not considered.
Strange results: naming in L1 in bilingual
context overactivated the network more
than in a monolingual context, but less
than naming L2.
BrainGlot 2009 - Barcelona
Language control and lexical competition in bilinguals: an
event-related fMRI study (Abutalebi et al., 2008)
Participants
12 early bilinguals (6 females). L1: Italian; L2: French; AOA=
less than 3. More exposed to L2 than to L1
fMRI task
Passively listening four types of sentences:
1. Control: L1
2. Control: L2
3. Language switch from L1 to L2.
4. Language switch from L2 to L1.
Results I (Abutalebi et al., 2007)
Switch vs non-switching activated the
LIFG, RIFG and bilateral superior temporal
gyrus.
BrainGlot 2009 - Barcelona
Results II (Abutalebi et al., 2008)
Forward
switching into a
less dominant
language (L1)
activated the
left striatum and
ventral ACC
Backward
switching into a
dominant
language (L2)
did not activate
the language
control network
BrainGlot 2009 - Barcelona
Discussion
Switching into a less dominant language
activated the language control network,
but not switching into a dominant
language.
This replicates Wang et al. study
BrainGlot 2009 - Barcelona
Study 1: Objective
To replicate Wang et al’s study in early
and high proficient bilinguals. Some
modifications were made:
ITI was 2 sec.
Performance was controlled: responses were
aloud
Cues and pictures were simultaneous-
BrainGlot 2009 - Barcelona
Picture Naming Task
Participants:19 Spanish/Catalan early bilinguals.
- 7 males and 12 females
- 11 bilinguals learnt one language at home and the other at
the kindergarten, and each father speaks a different language
in 8 participants
- High proficiency in both Spanish and Catalan.
- All subjects were right-handed.
- L1 and L2 were determined from infancy data: L1 was
Catalan for 11 and Spanish for 8.
BrainGlot 2009 - Barcelona
Task
Subjects were instructed to name simple pictures
in the correct language according to picture
colour (red, blue):
- Forty-four pictures of common objects with
non-cognate names
-Interstimuli interval: 2 seconds
-Switch and non-switch trials
-Responses were aloud.
BrainGlot 2009 - Barcelona
Naming sequence
COMETA / MILOTXA, ESTEL
LAVADORA / RENTADORA
CERDO / PORC
MANZANA / POMA
OJO / ULL
PATO / ÀNEC
BrainGlot 2009 - Barcelona
Task
There were 240 trials:
-120 switch trials: 60 Spanish Catalan
60 Catalan Spanish
-120 non-switch trials: 60 Spanish Spanish
60 Catalan Catalan
An examiner inside the scanner room registered
responses
BrainGlot 2009 - Barcelona
Behavioral results.
Omissions and commission errors
Percentage of errors was Catalan Spanish
%
low. 5
4
There were no significant
differences between 3
languages in omission 2
errors, but differences in 1
commission errors
reached significance (p C-C, I-I
Monolinguals suffered a greater
switching cost than bilinguals
This result has been also obtained
in children (Bialystok &
Viswanathan, 2009)
BrainGlot 2009 - Barcelona
Study 2: Objective
To investigate neural basis of task
switching in early and high proficient
bilinguals and monolinguals.
No previous studies on this topic, but task
swithcing has been associated with the
RIFG, the striatum and the ACC (Robbins,
2007).
BrainGlot 2009 - Barcelona
Task switching
Participants
19 Spanish/Catalan early bilinguals
- 11 bilinguals learnt one language at home and the
other in kindergarten, and in 8 each father speaks
a different language
- High proficiency in both Spanish and Catalan.
- All subjects were right-handed.
21 Spanish monolinguals: students from
monolingual regions just arrived to Castellon
BrainGlot 2009 - Barcelona
Task switching
-Scanner: 1.5 T Siemens
Task: Subjects were asked to press one of two
buttons according to a combination of geometrical
pictures
CUES: SHAPE OR COLOR. Cues simultaneously
presented to pictures.
BrainGlot 2009 - Barcelona
Task switching
COLOR COLOR SHAPE SHAPE SHAPE SHAPE COLOR COLOR
Thumb Index
button Button
BrainGlot 2009 - Barcelona
Task switching
Conditions:
-60 Non switch trials: Color-Color
or Shape-Shape
-60 Switch trials: Color-Shape or
Shape-Color
BrainGlot 2009 - Barcelona
Behavioral results
Switch costs: 32 ms for monolinguals and 4ms for bilinguals (p=0.051)
BrainGlot 2009 - Barcelona
Behavioral results
Switch costs: 4% for monolinguals and 0% for bilinguals (p Monolinguals: Increased activation in the left
inferior frontal gyrus (IFG) (blue cluster).
•Monolinguals> Bilinguals: Increased activation in the right IFG
(red cluster).
BrainGlot 2009 - Barcelona
Performance correlates in bilinguals
Lower switching costs Stronger activity in language control
areas (striatum and LIFG)
BrainGlot 2009 - Barcelona
Performance correlates in bilinguals
Higher switching costs Stronger activity in task switching areas
BrainGlot 2009 - Barcelona
Conclusions
- Monolinguals have shown significant switching costs and neural
activity in brain areas typically related to task switching: RIFG
and ACC.
-Bilinguals have shown no switching costs and a neural activity in
brain areas typically related to language control: LIFG
-Bilinguals seem to be an heterogeneous group: those with lower
switching costs activate the language control network (LIFG and
striatum), whereas those with higher switching costs activate as
monolinguals the rIFG.
The present results are consistent with the hypothesis that
bilinguals' early training in mixing languages leads to the
involvement of language control brain areas when performing
non-linguistic cognitive tasks.
BrainGlot 2009 - Barcelona
General Conclusions
Language swtiching in bilinguals activates
brain areas involved in language control:
LIFG, RIFG, striatum and ACC.
Language control in early bilinguals is more
related to the striatum
Language control in late bilinguals is more
related to ACC.
BrainGlot 2009 - Barcelona
General Conclusions
Set-switching in monolinguals is related to
rIFG.
Set-switching in early, high proficient
bilinguals is related to the language
control network: LIFG and the striatum.
Early and high proficient bilinguals may
overlap brain areas involved the language
and cognitive switching
Future studies should serve to delimitate
which factor or factors are responsible for
these effects
BrainGlot 2009 - Barcelona
Thank You very much....
BrainGlot 2009 - Barcelona