Psycholinguistics/Neural Basis of Multi-Lingualism

Introduction
Multilingualism is a skill possessed by individuals who can communicate in more than one language. This chapter introduces the neural structures that deal with multiple languages and discusses the factors affecting the neural organization within multilinguals. All people posses the ability to communicate in at least one language which is called their native language. These individuals are referred to as monolinguals and people who can communicate in more than one language are referred to as polyglots. The native language or commonly referred to as L1 is a bit tricky to distinguish in simultaneous polyglots who learned multiple languages at the same time during early development. However there is usually a slight preference for one language and for simplicity that language is designated the individuals L1, which then makes other subsequent languages L2, L3, etc. In successive polyglots assigning of L1, L2... is simple because each language learned subsequently gets assigned chronologically. A multitude of studies have been carried out to asses the neural organization within polyglots through neuroimaging techniques while subjects were administered different tasks. A few task and their subsequent neural activation shall be discussed later in this chapter. For an in depth analysis of single language localization within the brain please refer to previous chapters within this online textbook.

Factors Affecting Neural Organisation of Multilinguals
The neural organization of language is affected by multiple factors such as the age of L2 acquisition, the degree of proficiency attained in each language, the method L2 was taught, individual variability etc. Different methods of teaching L2 might favor different strategies for language processing, and hence distinct cerebral circuits. In terms of age of acquisition, multiple studies suggest that people who learn a language later in life do not generally achieve the same level of proficiency as native speakers (Birdsong, 1999; Johnson and Newport, 1989). The causes of these age effects in language performance vary widely from a biologically based critical period to the differences between infant and adult learning contexts (Lenneberg, 1967). For learners below the age of 18, age of exposure was strongly correlated with success in grammar attainment of L2 and no significant effect of aptitude level. However, young adults between ages 18 and 40 showed opposite results where aptitude was the main determinant in ultimate grammatical attainment rather than age of exposure. Furthermore, for the oldest learners over the age of 40 neither aptitudes nor age of arrival were good predictors of L2 grammar attainment (DeKeyser et al. 2009). The age of L2 acquisition seems critical for grammatical processing whereas lexical-semantic processing is less affected by age of acquisition but rather correlated with level of L2 proficiency (Wartenburger et al., 2003). Results from the imaging study showed no difference in brain activations between L1 and L2 during grammatical tasks for early and proficient bilinguals while highly proficient yet late L2 acquisitioned individuals showed additional neural activation. In lexical-semantic processing no differences were found irrespective of age of L2 acquisition, but great differences were found with varying proficiency (Wartenburger et al., 2003). Another interesting find using voxel-based morphometry is that early acquisition highly proficient subjects have increased grey matter density in the inferior parietal cortex, which is the area associated with L2 proficiency. These results suggest that language talent in polyglots resides within the left inferior parietal lobe (Mechelli et al., 2004).

Critical Period Hypothesis
Performance results in L2 of adults, even experienced L2 users, hints that the ability of effectively acquiring an L2 is limited by a critical period. The critical period hypothesis suggested by Lenneberg (1967) is dependent on the assumption that age related effects in L2 studies are a consequence of changes in brain structures due to maturation and these structures are critical in the learning and processing of language. The notion of a sensitive period for language acquisition comes from the loss of flexibility for cerebral reorganization after puberty (Scovel, 1988; Patkowski, 1980, 1990; Sakai, 2005). Other studies have proposed that age-related changes in L2 performance arise from the nature and extent of the interaction between L1 and L2 (Oyama, 1979; Flege, 1987, 1988, 1995, 1998b; Bialystok, 1997). This approach assumes that the more developed an L1 system is when L2 is being learned, the more influence L1 will have on L2 learning. The debate between maturation and interaction however are still unresolved and as suggested, neural development and L1 acquisition during childhood is extremely complex (Bates & Goodman, 1998). The length and duration of the critical period for L2 acquisition also has not been successfully established. Some argue that it ends at 12 years of age (Scovel, 1988), whereas according to others (Patkowski, 1990) it ends at age 15. Lenneberg (1967) suggests that puberty is the end of the critical period which is problematic since it is highly variable between individuals. Successful language acquisition requires different linguistic abilities which develop according to their own developmental timeline, therefore in addition to an acquisition critical period, linguistic ability may also have its own critical period, and both these critical periods may have to overlap in order to attain native-like proficiency in L2 (Sakai, 2005).

L2 acquisition after the critical period is different in nature from critical period learning, because of the requirement for a more conscious, labored effort. Studies have also shown a slow systematic decline in macroscopic levels of brain metabolism as well as synaptic density between the ages of 4 and 20 (Bates et al., 1992). Assuming there is a critical period for language acquisition, late L2 acquisition will likely resemble L1 acquisition because the process will be governed by the same neural structure. Therefore, L2 acquisition during the critical period should have little or no influence from L1 because of similar access to the construction of a language system. Learning L2 after the critical period however would reflect elements of L1 because available resources would be activated and they would begin with the linguistic structures already in place (Bialystok & Hakuta, 1999). In contrast another theory claims that the processing of L2 acquired late in life depends upon different cognitive mechanisms and cerebral structures from L1 (Ullman, 2001). In this regard L2 grammatical knowledge is declarative while L1 is implicit whereas both L1 and L2 are declarative for lexical knowledge. Since implicit and declarative knowledge are mediated by distinct neural systems, left frontal-basal ganglia and left temporal areas respectively, this differential hypothesis claims that late L2 acquisition cannot depend on the same brain mechanisms used for L1 (Ullman, 2001).

Neural anatomy in language
Macroscopic brain functional neuroimaging studies have shown that highly proficient bilinguals activate the same set of brain regions irrespective of language (Perani, 2005). This suggests highly overlapping and interconnected neural circuits, but do not indicate how the brain controls or determines which language to use. Anatomically the bilingual brain cannot be summarized as the sum of two monolingual language systems, but rather should be represented as a unique and complex neural system that differs in each individual. Monolinguals appear to consistently activate relatively fixed left-hemispheric network groups, but polygots with L2 learned later in life and with less attained proficiency activate non-overlapping and less consistent cerebral areas. However, when proﬁciency was kept constant, age of acquisition did not seem to have a major impact on brain representations of L2 (Abutalebi et al., 2001). The cerebral network for reading should operate differently between the L1 and L2 languages since it has been shown that even proﬁcient bilinguals are slower to recognize L2 than L1 (Grainger & Beauvillain 1987; Jiang, 1999; Thomas & Allport, 2000). It is inconclusive whether and to what extent the bilingual brain can ultimately decode L2 automatically without effortful cognitive control. It has been shown that even before processing lexical information, bilinguals filter out the non-target language (Rodriguez-Fornells et al. 2002).

This control is speculated to occur within the left occipitotemporal cortex since this region is sensitive to language specific orthographic patterns and has control over an early and unconscious stage of reading (Dehaene et al., 2001; Nakamura et al. 2005; Vinckier et al. 2007). Voluntary regulation is needed to control the left inferolateral frontal cortex involved in switching between languages for speech production and reading aloud (Fabbro et al., 2000; Hernandez et al., 2001 ; Price et al., 1999). The left occipito-temporal cortex is thought to house semantic level representations of the languages and therefore may exhibit differential sensitivity to L2 (Chertkow et al., 1997; Devlin et al., 2004; Nakamura et al., 2007). Evidence exists to support a general purpose cognitive control system in the inferiolateral prefrontal cortex to resolve conﬂict between languages by enhancing neural representations of task-relevant information in posterior temporal cortex based on the language in use (Egner & Hirsch 2005). Multiple neuropsychological and imaging studies have failed to establish consistent neuronal substrates for L2 acquisition, perhaps because they were obtained in various languages, differences between languages, differences in tasks, subject proficiency variability and also inter subject variability.

Verbal Fluency Tasks


Verbal ﬂuency tasks showed consistent brain activity involved in the left lateral prefrontal cortex, and in particular, the inferior and middle frontal gyrus. Additional brain activity in the left hemisphere was located in the premotor cortex, the inferior parietal lobe, the superior and middle temporal gyri and the thalamus. A few right hemispheric brain activations were also found in the inferior frontal gyrus and right cerebellum. Production of L2 words caused significant activations in the left inferior and middle frontal gyri, left premotor cortex, and insula. Further activity was also observed in the left inferior parietal lobe, left caudate nucleus and right inferior frontal gyrus. A more concentrated activation was observed in L1 word production tasks within the left frontal lobe, specifically in the inferior and middle frontal gyri and insula (Perani et al., 2003).

Word Generation Tasks
Word generation tasks of increasing task difficulty showed brain activity in Brodmann areas 45, 46, and 47 in the left inferior and middle frontal gyrus. These areas are associated with the selection processes of information among competing alternatives from semantic memory. These cortical areas would be specifically activated on the basis of task difficulty, when the word-to-be-generated possessed more competing alternatives a larger left prefrontal network was involved (Thompson-Schill et al., 1997 ; 1999). In word generation or other production tasks, there is evidence that a lower degree of language proﬁciency may be associated with differences in brain activity in anterior brain structures, such as Broca’s area and the basal ganglia (Chee et al., 1999; Yetkin et al., 1996). In comprehension tasks however, the temporal lobe was associated with proﬁciency-related differences (Dehaene et al., 1997; Perani et al., 1996, 1998). More extensive cerebral activations are associated with production of the less proﬁcient language and smaller activations when comprehending the less proﬁcient language. In effortful tasks such as word generation, the difference in activation can be attributed to the recruitment of additional resources. In language comprehension, the less extensive activation associated with the less proﬁcient language may reﬂect a more limited knowledge of the linguistic material (Abutalebi et al., 2001).

Reading tasks


During reading tasks cerebral networks showed differential activation between L1 and L2 but common brain regions such as the bilateral occipitotemporal cortex, bilateral medial frontal areas and the left inferior parietal and left lateral prefrontal areas were activated in both languages (Nakamura et al., 2010). These results support the dual-route connection hypothesis stating that words between L1 and L2 are directly linked at the lexical level and indirectly linked at the semantic level. Further neuropsychological evidence shows that the left lateral temporal cortex is involved in lexicosemantic processing during reading tasks (Chertkow et al., 1997). The medial temporal gyrus is thought to be involved in lexical activation since a region just posterior to it shows increasingly greater response as letter strings become more like real words (Vinckier et al. 2007). The language non-selective frontal gyrus-medial temporal gyrus stream probably mediates lexical activation in each language but not interlanguage lexical activation or semantic activation. Postlexical mapping of word-form to meaning in the left medial temporal gyrus is much less automatic for L2 and compensated by an effortful cognitive control for amplifying the lexicosemantic association between each language (Nakamura et al., 2010).

Mental Counting Tasks
Mental counting tasks produced similar activations for all L1 languages with prominent left frontal activation in Broca’s area and medial frontal or cingulate areas. The cingulate areas concurrently activated the left superior parietal lobule and a small right anterior region was. In addition to these activations L2 languages also activated posterior left temporal regions (Vingerhoets et al., 2003). In mental picture naming tasks superior medial frontal and extensive bilateral occipital activation was observed for all L1 languages, whereas L2 languages additionally activated left frontal regions including Broca’s area and left superior parietal and temporal occipital areas (Vingerhoets et al., 2003). Mental language production experiments found Broca’s area differentially activated for early vs late bilinguals where similar activations were seen for L1 and L2 in early learners, and spatially segregated activations in late learners (Kim et al. 1997).

Verb Generation Tasks
Verb generation tasks produce reduced brain activity in the left frontal lobe following practice mainly related to processing differences between high and low practice performance. The reduction in prefrontal activation after repeated practice suggests a decreased dependence on controlled processing. In terms of language processing, increased practice does not correlate to increased proficiency but rather a greater exposure to the language being used. Intensive exposure to a language in a bilingual environment will lead to a lower activation threshold and a higher degree of automaticity, thereby decreased dependence on controlled and attentional processing (Green, 1986). In tense generation tasks cortical plasticity for L2 gets increasingly similar to L1 specialization in the inferior lateral gyrus despite noticeable differences in their performance and linguistic knowledge between L1 and L2. If proficiency level was higher however, the activation in the left dorsal inferior frontal gyrus was lower, indicating that proficiency plays a large role in activation of this area (Sakai, 2005). This finding makes sense because during early acquisition of L2 when proficiency is low, the cortical regions have to be strongly elicited, whereas when proficiency gets high enough the process becomes automated reducing the need for activating the left dorsal inferior frontal gyrus.

Listening Tasks
Listening tasks to L1, showed remarkable consistency in the observed activated areas in the left hemisphere. All subjects showed activity in the left temporal lobe along the superior temporal sulcus, as well as portions of the superior and middle temporal gyri. Some showed activation in the temporal pole and others the left angular gyrus. Although similar activity was occasionally found in the right temporal lobe, including the right superior temporal sulcus and temporal pole, it was always weaker, highly variable from subject to subject, and never extended into the right angular gyrus. Outside the temporal lobe, the only consistent focus of activation was found near the intersection of the inferior frontal sulcus and the precentral sulcus around Brodman’s areas 6, 8, 9 and 44 (Dehaene et al., 1997). Listening to L2 showed much greater inter-subject variability and there was no significant similarity in anatomically activated areas between the subjects. In a few subjects there were some activations similarity to L1 in the same areas of the left temporal lobe but much more dispersed in comparison. No activity was found in the left temporal pole and angular gyrus when listening to L2. When subjects did show left temporal activity in L2, the activation volume was smaller than in L1, and activated additional small subregions in the right temporal lobe mostly the right superior temporal gyrus and sulcus (Dehaene et al., 1997). Cerebral activations outside the temporal lobe was also observed while listening to L2 but were highly variable. Some showed a highly speciﬁc activation of the left inferior frontal gyrus (Broca’s area) and of the inferior precental sulcus in L2, but not in L1. Others showed activity in the left and right anterior cingulate when listening to L2, but not L1. Yet another few showed frontal activation at a region similarly activated in L1, the intersection of the inferior frontal sulcus and the precentral gyrus. (Dehaene et al., 1997). The anterior cingulate region is known to be active in attentive, controlled or processing tasks implicating that the subjects had to engage greater attentional resources for processing L2 compared to the more automatized L1. These results indicate that L1 acquisition relies on a dedicated left-hemispheric cerebral network, while late L2 acquisition is associated with variable cortical substrates based on individual differences.

Priming Tasks
‎During priming tasks, irrespective of language, semantically related prime and target words reduce the activation of the left anterior temporal cortex. This finding suggests that different languages use the same neuronal populations within the temporal cortex rather than activating different neurons within the same region. However, in the head of the left caudate, only semantically related words of similar language were primed and therefore reduced activation. Two proposed mechanisms for these results have been suggested with the first being: within this neural region different languages are processed by language selective neural populations and for the prime language, semantic priming selectively adapts the activations of neuronal populations. There was no reduction in activation of semantic primes between different languages because the change in language activated a different neuronal population (Grill-Spector et al., 2006). The second proposed mechanism is that the same neuronal populations respond to similar semantic primes but increase firing when the languages are different. This indicates that the neuronal populations have the ability to distinguish difference between similar or distinct language primes and manipulate their firing accordingly (Naatanen et al., 1997). Alternative studies have suggest that the left caudate is required to monitor and control lexical as well as multilanguage synonym production. The highest responses in the left caudate are seen when there is a change in language or a change in meaning but lowest when the context of words are related in both language and meaning. Morphologically, the left caudate receives projections from frontal, temporal, and parietal association regions in the language-dominant hemisphere connections back to these areas via the thalamus. These connections play a critical role in controlling and selecting automatic motor sequences such as those necessary for articulation. A mechanism sensitive to the language being used is needed since the motor patterns differ across languages and the head of the left caudate may be ideally situated to serve this function. Functional neuroimaging studies have demonstrated that neuronal responses in the head of the left caudate are sensitive to both the semantic content and the language of written words (Crinion et al., 2006). The left caudate is crucial in language control so further research involving the cortical structures connected to the left caudate is critical in the progress of understanding neural control of multiple languages. Unconscious neural priming is language non selective during bilingual word recognition in the left midfusiform gyrus but exhibits a preference for L1 in the left posterior middle temporal gyrus (Nakamura et al., 2010).

Summary
Multiple factors influence the neural organization of language within an individual such as age of acquisition, proficiency of the language, individual variability, etc. Successful language acquisition seems to be restricted by a certain critical period early on in development but the exact range of the critical period has not yet been determined. Attainment of native like proficiency in an L2 language is restricted by this critical period, but L2 can be learned at anytime throughout life. Learning a language later on in life however is more labor intensive as more neural resources are recruited to translate L2 to L1 in order to comprehend. Additionally, a person who learned a new language post critical period already has a language established and therefore the new language will be developed in comparison to the native language. This can cause a lot of obstacles because of the grave differences between many languages in terms of grammar and structure. Overall language is localized to the left hemisphere of the cerebral cortex. Highly proficient multilinguals who acquired all their languages within the critical period and are equally proficient in all their languages activate the similar neuronal populations when using either language. However a less proficient multilingual activates the same neuronal population as a proficient individual but recruits additional neural populations depending on the task. As an individuals gets more proficient in a language the neural activations decrease and more closely resemble the activations of a native speaker. However, even a highly proficient individual will never depict the same activations as a native speaker if the language was acquired after the critical period.

Essay Questions
Some possible questions to facilitate further thought and insight into the topic of Multilingualism.

1) Discuss some of the similarities and differences between simultaneous and successive multilinguals if age is kept constant (i.e. learned during the critical period).

2) Discuss some of the differences between similar cortical insults in early vs late language acquisitioned individuals.

3) In terms of speech problems, will therapy in one language also have an impact on the other languages?

4) Describe the progression of neural activations during each task and justify the sequence.
 * 1) A reading task in L2
 * 2) A verbal production task in L2
 * 3) A listening task in L2

5) Discuss some advantages with learning language during the critical period based on the following areas.
 * 1) Neural real estate
 * 2) Conscious effort
 * 3) Native like fluency

6) Based on evidence from development studies and language acquisition experiments outlined in the chapter, discuss your own thoughts on the critical period effect and its effect on language acquisition.

7) How important is bilingual parental involvement? Discuss some issues regarding amount of exposure to different languages during early development as well as some advantages and possible obstacles for children with multilingual parents.

8) Discuss the role of a possible language acquisition device and any constrains it might have based on the critical period.

9) Discuss the effect of the maturational state of the brain and its influence on language acquisition. (i.e. neural differences between children and adults and any influences these might have on language acquisition)

10) Suppose Jim was born to bilingual parents and exposed to French and English equally. Jim however resides in Spain and therefore when he starts school will be exposed to mainly Spanish. Discuss some of the obstacles and advantages Jim might face by being a bilingual prior to staring school.

11) Suppose an individual is completely fluent in his/her native language prior to entering school. Once in school, the child learn a second language fluently and reaches native like proficiency due to learning during the critical period. Now assume that L2 is mainly used during many years of the individuals life, and exposure to L1 is seldom. Is it possible that L1 could be completely lost after many years? Also if it is lost, would the person be able to achieve native-like proficiency if they were to be heavily exposed to the language again after the critical period has ended?

Quiz
{ Language acquisition in humans is constrained by a { critical period }.
 * type="{}"}

{ According to Lenneberg, the critical period threshold for language acquisition in humans is probably around { puberty }.
 * type="{}"}

{ According to Wartenburger, { grammatical } processing is highly dependent on age of L2 acquisition, whereas { lexical-semantic } processing is dependent on L2 proficiency.
 * type="{}"}

{ After the critical period, heavy { exposure } is the only way to attain good proficiency in an L2, but { native-like } proficiency will never be achieved.
 * type="{}"}

{ The occipitotemporal cortex is sensitive to language specific { orthographic } patterns and controls { unconscious reading }.
 * type="{}"}

{ Monolingual language activation is mostly { left } hemisphere { lateralized }. { During performance tasks, { proficient } polyglots activate { identical } neural populations regardless of which language they use. { In late acquisitioned polyglots, during tasks in { L2 }, additional { neural populations } are recruited over populations already activated by the { native language }.
 * type="{}"}
 * type="{}"}
 * type="{}"}

{ Verbal fluency tasks activate the left { prefrontal cortex } and { frontal gyrus } in both monolinguals and polyglots. { In lower proficient individuals during a production task, there are differences in the activation of { Broca's Area } and the { basal ganglia } when compared to native speakers.
 * type="{}"}
 * type="{}"}

{ The recruitment of additional neural resources in the { prefrontal cortex } indicates that lower proficient bilinguals process secondary languages { effort-fully or consciously }.
 * type="{}"}

{ During reading tasks, the left temporal cortex is involved in { lexicosemantic } processing.
 * type="{}"}

{ { Semantic } primes across different known languages reduced the activation in the left { temporal } cortex of proficient polyglots.
 * type="{}"}

{ Highly proficient polyglots who acquired all their languages prior to the end of the { critical period } showed little { detriment } to all acquired language.
 * type="{}"}

{ The neural activations of an individual gains closer resemblance to that of a { native } speaker when { proficiency } in that language is { increased }.
 * type="{}"}

Deeper Understanding
To take things further in your understanding of neural basis of multilingualism, exposure to further research should be initiated. This can be done either by further literature searches or better yet, conducting your own experiments.

Here is some large scale experiment questions that could be answered with the right resources. Considering most of the experiments done on multilingualism compare Polyglots to controls and include brain imaging studies, the next step would be to do some testing on multilingual aphasics. In addition to various performance tasks, comprehension and learning tasks can be measured to see any detriments to either language possessed. Apart for language acquisition detriments from aphasia, research on age or gender has not been extensively examined. As a result comparable cohorts of age and gender can be tested on language acquisition and performance to see if any significant differences exist. To initiate thought on this matter, would sexual dimorphic differences in brain structures or hormonal levels play any significant difference in language acquisition or performance on specific language tasks.

Here is a small scale experiment for someone to get a deeper understanding of multilingualism acquisition in the human brain. Generate a short list of words in various different branching languages. Attain reasonably sized cohorts of various age groups such as children, adolescents, adults and seniors. Furthermore for comparison, replicate the same cohorts but with individuals who are early acquisitioned polyglots. As a simple test ask all the groups to memorize and recall the word lists when restricted to the same amount of time. From the understanding of the information outline in the chapter, which group would you expect to have the highest average score on the memory recall tests and explain why? As a more complex test, while the cohorts are memorizing the word list, make a distraction task that keeps various brain regions occupied like the temporal lobe, pre-frontal cortex, etc. Now depending on the regions that are preoccupied, which cohort would you expect to do the best on the memory recall of different language words and why?