PET studies on the memory processing of word pairs in bilingual Finnish–English subjects

Behavioural Brain Research 132 (2002) 47 – 57 www.elsevier.com/locate/bbr

Research report

PET studies on the memory processing of word pairs in bilingual Finnish –English subjects U. Halsband a,*, B.J. Krause b,c, H. Sipila¨ d, M. Tera¨s d, A. Laihinen d a

Department of Psychology, Neuropsychology, Albert-Ludwigs-Uni6ersity of Freiburg, Niemensstraße 10, D-79098 Freiburg, i. Br. Germany b Department of Nuclear Medicine, Heinrich-Heine-Uni6ersity of Du¨sseldorf, Du¨sseldorf, Germany c Department of Nuclear Medicine (KME), Research Centre Ju¨lich, Ju¨lich, Germany d PET Centre, Uni6ersity of Turku, Turku, Finland Received 20 April 2001; received in revised form 10 September 2001; accepted 10 September 2001

Abstract This study examined the fundamental question whether verbal memory processing in two unrelated languages is mediated by a common neural system or by distinct cortical areas. Ten right-handed, male Finnish– English adult late bilinguals who had acquired the second language after the age of 10 were scanned whilst either encoding/retrieving word pairs in their mother tongue (Finnish) or in a foreign language (English). Within each language, subjects had to encode and retrieve four sets of 12 visually presented paired word associates which were not semantically related. Two sets consisted of highly imageable words (e.g. monkey-table; koira-lasi) and the other two sets of abstract word pairs (e.g. freedom-moral; uhka-suure). Presentation of pseudowords served as a reference condition. An emission scan was recorded after each intravenous administration of O-15 water. Encoding was associated with prefrontal and hippocampal activation. During memory retrieval, precuneus showed a consistent activation in both languages and for both highly imageable and abstract words. Although the brain mechanisms of the two languages share common components, differential activations were found in Broca’s area and in the cerebellum as well as in the angular/supramarginal gyri according to the language used. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Bilingualism; Verbal memory; Associative learning; PET; Precuneus; Prefrontal cortex

1. Introduction Recent studies have shown variations in the cerebral activation in the context of processing native and foreign languages [11,34,35,59,60]. It has been suggested that there are differences in the cerebral organization of foreign language depending on the age of acquisition and learning strategies [50,51,80]. It is to a large extent unsettled how the multilingual processing takes place in the brain. * Corresponding author. Tel.: + 49-761-2032475/+ 49-1729795000 (Mobile); fax: +49-761-2032417. E-mail address: [email protected] (U. Halsband).

Clinical studies on aphasic disturbances in bilingual patients have provided information about the recovery of the respective languages spoken by these subjects, e.g. preferential recovery of the old as contrasted with the new language or preferential recovery of the most familiar language [14,58,68]. These findings suggest that there are multilingual patients, who after brain lesions may become aphasic in only one of the languages they originally mastered. This dissociation is supported by the results obtained with electrical cortical stimulation: Ojemann and Whitaker [53,54] stimulated electrically the cortical areas of neurosurgical patients in order to identify the speech-relevant centers. In the study by Ojemann & Whitaker (1978) the localization of two languages in the lateral cortex of the dominant cerebral

0166-4328/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 4 3 2 8 ( 0 1 ) 0 0 3 8 6 - 2

48

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

hemisphere was determined by the technique of mapping sites where electrical stimulation altered naming in two biligual patients’ languages (patient 1 Dutch–English, patient 2 Spanish– English). The patients had to name 45 common objects which were presented visually through slides. It was found that sites in the centre of the language area of each patient were involved in both languages. However, peripheral to this, in both frontal and parietal cortex, were sites involved in only one of the languages. It was concluded that, in each patient, each language in part used different areas of the brain. A recent study with optical imaging preceding a neurosurgical procedure [64] confirms that cortical language representations in bilingual persons may consist of both overlapping and distinct components. Using functional magnetic resonance imaging (fMRI), Kim et al. [34] reported that the mother tongue is localized in Broca’s area and each newly acquired language in anterior portions of Broca’s area. The second language tended to have a more diffuse representation in the left hemisphere than did the mother tongue [11]. Furthermore, it has been proposed that the left supramarginal gyrus in the parietal lobe controls switching from one language to another [66]. Functional imaging studies focussing on one language at a time have provided evidence for a distributed set of brain areas subserving verbal memory with a distinction between regions, which are activated both in encoding and retrieval on one hand (prefrontal cortex, anterior cingulate cortex), and others showing task-related activations (retrieval: precuneus) [16,17,23, 29,32,38,39,49,73,75,78]. During verbal memory retrieval, the predominantly right dorsolateral prefrontal cortex, the anterior cingulate and the precuneus showed consistent activation. Precuneus activation occurred in visual and auditory presentation modalities and for both highly imageable and abstract words [39]. In the present investigation, our aim was to use O-15 water PET to study differences in cerebral activation patterns associated with the verbal memory processing of concrete and abstract word pairs in the native language of the subjects (Finnish) compared with a fluent foreign language (English). Therefore, the aim of the present study was to disentangle the neural mechanisms of encoding and retrieval in bilingual subjects who were fluent in two widely different linguistic groups, i.e. a non-Indo-European, Finnish and an Indo-European language, English. We looked for shared and nonshared neural substrates in a paired-word association paradigm for such contrasting languages. This is the first study to demonstrate that the precuneus shows consistent activation during memory retrieval of word pairs from two structurally very distant linguistic groups using both highly imaginable and abstract words.

2. Subjects and methods

2.1. Subjects Ten late acquisition bilinguals (i.e. English was a language acquired at school after the age of 10 years) took part in this study. They were right-handed male subjects (age 27.39 5.1 years, mean9SD, range 22–36 years) with no known history of neurological or psychiatric illness. Their mother tongue was Finnish and they had learnt 3–4 languages at school, the first and predominant of these was English. Our subjects reported that they were confronted with English in their daily activities (work and relaxation including music and films or television programs). All but one of the subjects had undergone an MRI examination (1.5 T) of the brain, where no structural abnormalities were found. The one subject, who could not take an MRI examination due to metal implants in his denture, had an unremarkable computerized tomography (CT) scan. The research project was approved by the Joint Ethical Committee of the University and University Hospital of Turku. Each subject gave informed, written consent for participation in the study according to the guide-lines of the Declaration of Helsinki.

2.2. Positron emission tomography (PET) imaging O-15-labelled water was produced with a low-energy deuteron accelerator Cyclone 3 (Ion Beam Application, Inc., Louvain-la-Neuve, Belgium). Cyclone 3 is a compact cyclotron for hospital use, and it accelerates positively charged deuteron ions up to 3.8 MeV to generate O-15-labelled oxygen gas for PET applications. This gas was processed to water vapour with hydrogen gas on the palladium catalyst. Water vapour was mixed with sterile saline using dialysis technique in a continuously working water module [8]. The half-life of O-15 is 123 s. Sterility and pyrogen tests were performed to verify the purity of the product. Scans of rCBF were obtained for each single subject using a GE Advance PET Scanner (General Motors Medical Systems, Milwaukee, Wisconsin, US). This apparatus has been previously described [43]. It has 18 detector rings with 672 crystals/ring (6× 6 blocks) and provides 35 transverse sections through the brain spaced 4.25 mm apart (center to center/axial sampling interval) covering 152 mm axially (axial field of view) with an aperture of 550 mm. Transmission scans performed with a Ge-68/Ga-68 source were used for measured attenuation correction. The head of the patient was placed correctly with a laser-positioning system according to the cantho-meatal reference line. A filtered back-projection algorithm was employed for image reconstruction, on a 128×128 matrix. The spatial resolution full-width half-maximum (FWHM) was 8 mm.

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

Regional cerebral blood flow (rCBF) during each cognitive task was measured by recording the distribution of radioactivity in the brain following an intravenous injection of 200 MBq of O-15-labelled water (10 ml in 10–15 s) through a forearm cannula. Twelve tasks were carried out during a 3-h session. The minimum interval between the O-15 water injections was 10 min. As each of the 12 scans was concerned, the cognitive task began 15 s before the administration of O-15 water. Emission data were acquired in 3-dimensional mode for 90 s starting at the entry of the tracer into the brain, for which the criterion was the true coincidence rate exceeding the threshold of 15 000 counts/s. The data were framed into a single static frame of 90 s [30,41].

2.3. Experimental paradigm A verbal memory task was used. Subjects were visually presented word pairs. The subjects had to encode and retrieve four sets of 12 visually presented paired word associates. Study words were two-syllable Finnish or English words. The word pairs were semantically unrelated and, therefore, difficult to associate. ‘Hard’ word associations (as introduced by Wechsler [81] in his Wechsler Memory Scale, Subtest VII) were used to increase the mnestic demands. We used a within-subject design, which meant that each subject underwent both a native language session and an English session. The stimulus words were presented on a 21¦ computer screen placed at a distance of about 70 cm from the eyes (Font: Times New Roman, size: 72 points). Subjects were instructed to read them aloud (duration of presentation 4 s, 1 s interval) and to learn the paired associations. In order to avoid lateralization effects the second word was written under the first word, the letters were black on a white screen and centered. Between encoding and retrieval scans, the same word pair associates were presented in random order 1–3 times according to the number of encoding repetitions needed to retrieve at least 80% of the word pair associates. During retrieval scans, the first word of the pair was shown. The subjects had to read the first word aloud and to retrieve the associated word from memory and express it verbally. The reference task was the presentation of two-syllable non-sense words formed according to the spelling rules, but having no semantical meaning. The subjects were required to read the non-sense words aloud, but they were not requested to memorize them. The details of the experimental paradigm of pairedword association memory have been described earlier [29,38,39]. In the present study scanning was done on 2 separate days. In the first scanning session the subjects were requested to learn and to retrieve word pairs in

49

their mother tongue; the second scanning session involved paired-word association learning in the foreign language.

2.3.1. Session I (Finnish) Each subject underwent 12 O-15 water PET scans. Two word pair sets consisted of highly imageable words (e.g. POLKU — small road vs. KUVA— picture) and another two sets of abstract words (e.g. UHKA— threat vs. SUURE—unit). These words were selected on the basis of the Finnish word frequency handbook [70]. 2.3.2. Session II (English) The same subjects were scanned during the learning and retrieval of English word pairs. The English PET session was carried out on a different day after session I. The technical execution of the PET scans and the principles of the experimental paradigm were the same as described above, except for the fact that the word pairs were in English. The English proficiency of the subjects was tested with an ad hoc questionnaire before starting the experimental PET sessions. The subjects had to encode and retrieve four sets of 12 English word pairs. Word pair associates were presented visually. Two sets consisted of highly imageable words (e.g. MONKEY vs. BUTTER) and another two sets of abstract words (e.g. FREEDOM vs. POWER). The English words were chosen according to Paivio et al. [55,56], independently of the selection of the Finnish stimulus words. In order to rule out familiarity effects, great care was taken that the same nouns were used twice in the two different languages. For instance, if in the English condition the high frequency word ‘monkey’ was used, the Finnish word list did not contain a noun referring to this specific animal category or a similar species (apes); instead the Finnish word list made use of the word ‘koira’ (dog). 2.4. Analysis of data The data were first transformed into the ANALYZE format using a converter program especially developed for this purpose at Turku PET Centre. The actual quantitative analysis of the 90 s images was carried out with Statistical Parametric Mapping (SPM96, The Wellcome Department of Cognitive Neurology, London, UK) software [19,20] on a SPARC 20 workstation (Sun Microsystems). Calculations were performed with Matlab version 4.2c. Each reconstructed O-15 water scan was realigned according to the bi-commissural line into a stereotaxic space corresponding to the atlas of Talairach and Tournoux [76] using a PET template and normalized according to Friston et al. [19]. A Gaussian filter with a full width half maximum (15 mm) was applied to smooth each image to compensate for inter-

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

50

0.0001) and at least belonged to a cluster of 33 activated voxels (PB 0.05, corrected for multiple comparisons) [18].

subject differences and to suppress high frequency noise in the images. Differences in global activity within and between subjects were removed by the analysis of covariance (ANCOVA) on a voxel by voxel basis with global counts as covariate of regional activity across subjects for each task, as inter- and intrasubject differences in global activity may obscure regional alterations in activity following cognitive stimulation. For each pixel in stereotactic space the ANCOVA generated a condition-specific, adjusted mean rCBF value (normalized to 50/100 ml/min) and an associated adjusted error variance. The ANCOVA allowed comparison of the means across the different conditions using t statistics. The resulting map of t values constituted a statistical parametric map [19]. The results were analyzed both using a comparison to the reference task and on the other hand, subtracting the foreign language from the mother tongue and vice versa (cognitive subtraction). Voxels were identified as significantly activated if they passed the height threshold of Z= 3.72 (PB

3. Results

3.1. Foreign language 3.1.1. Encoding of high imagery word pairs Encoding of high imagery word pairs compared with reference task: Table 1 shows that the prefrontal cortex (BA 10) was activated on both the sides (right: Z= 4.46, left Z= 3.90), whereas the orbitofrontal cortex (BA 11) was significantly active only on the left side (Z= 4.03). The hippocampal formation was bilaterally activated (right: Z= 3.44, left: Z= 4.50). Furthermore, the angular/supramarginal gyrus (BA 39) was active on the left side (Z=3.95) and the precuneus (BA 7) on the right side (Z= 3.68).

Table 1 Encoding of words with high imagery content compared to reference condition Talairach coordinates

Z

BA

Region

Prefrontal cortex (right) Orbitofrontal cortex (left) Prefrontal cortex (left) Hippocampal formation (right) Hippocampal formation (left) Angular/supramarginal gyrus (left) Precuneus (right)

x

y

z

Foreign language 36 −22 −26 20 −14 −46 6

42 26 54 −38 −38 −72 −78

36 0 12 −28 −20 28 52

4.46 4.03 3.90 3.44 4.50 3.95 3.68

10 11 10

Mother tongue 8 32 18

26 −58 −36

−24 −4 −12

3.80 3.44 4.59

11

39 7

Orbitofrontal cortex (right) Hippocampal formation (right) Hippocampal formation (left)

Table 2 Encoding of words with low imagery content compared to reference condition Talairach coordinates

Z

BA

Region

x

y

z

Foreign language 36 −20 −4 4

46 48 −38 −78

32 0 28 56

4.37 3.72 4.08 4.02

9 10 31 7

Prefrontal cortex (right) Prefrontal cortex (left) Cingulate gyrus (left) Precuneus (right)

Mother tongue 18 −48 0

62 16 −62

4 −12 −28

3.24 3.50 3.53

10 47

Prefrontal cortex (right) Prefrontal cortex (left) Cerebellum, vermis

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

51

Fig. 1. Foreign language compared against reference condition. Comparison of adjusted mean rCBF in 10 subjects between retrieval of word pairs and reference. Spatial distributions of significant voxels are shown as integrated projections along sagittal, coronal and transverse views of the brain: (a) retrieval of high imagery word pairs; (b) retrieval of low imagery word pairs. Fig. 2. Mother tongue compared against reference condition. Comparison of adjusted mean rCBF in 10 subjects between retrieval of word pairs and reference. Spatial distributions of significant voxels are shown as integrated projections along sagittal, coronal and transverse views of the brain: (a) retrieval of high imagery word pairs; (b) retrieval of low imagery word pairs.

Fig. 3. Cognitive subtraction ‘foreign projections along sagittal, coronal and Fig. 4. Cognitive subtraction ‘mother projections along sagittal, coronal and

language minus mother tongue’. transverse views of the brain: (a) tongue minus foreign language’. transverse views of the brain: (a)

Spatial distributions of significant voxels are shown as integrated retrieval of high imagery words; (b) retrieval of low imagery words. Spatial distributions of significant voxels are shown as integrated retrieval of high imagery words; (b) retrieval of low imagery words.

Table 3 Retrieval of words with high imagery content compared to reference condition Talairach coordinates

Z

BA

Region

x

y

z

Foreign language 18 14 −22 60 0

54 16 48 −54 −78

0 −16 0 28 44

6.77 4.03 4.54 4.83 6.35

10 11 10/11 39/40 7

Prefrontal cortex (right) Prefrontal cortex (right) Prefrontal cortex (left) Angular/supramarginal gyrus (right) Precuneus (bilateral)

Mother tongue 20 −14 −4 12 −8 −12

54 56 −14 −72 −76 −36

−4 −12 20 36 48 −20

5.65 4.75 5.50 3.94 4.84 4.93

10 10/11 25 7 7

Prefrontal cortex (right) Prefrontal cortex (left) Cingulate gyrus (left) Precuneus (right) Precuneus (left) Cerebellum (left)

3.1.2. Retrie6al of high imagery word pairs Retrie6al of high imagery word pairs compared to reference task: As shown in Fig. 1a significant increases

in rCBF were seen in the precuneus (BA 7: bilateral Z= 6.35) and in the prefrontal cortex (BA 10: right Z= 6.77, BA 10/11: left Z=4.54). Furthermore, there

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

52

ti6e language’ (Table 5 and Fig. 3b) indicated activation of the left-sided medial frontal gyrus (BA 8, Z=5.41).

was a significant increase in rCBF (Table 3) in the right angular and supramarginal gyri (BA 39/40: Z = 4.83). The cogniti6e subtraction ‘foreign language minus mother tongue’ just passed the height threshold and showed (Table 5 and Fig. 4a) right-sided activation of the lower medial temporal area (BA 20, Z = 3.76).

3.2. Nati6e language 3.2.1. Encoding of high imagery word pairs Encoding of high imagery word pairs compared with the reference task: As shown in Table 1 the orbitofrontal cortex (BA 11) was activated on the right side (Z= 3.80) and the hippocampal formation on both the sides (right: Z=3.44, left: Z= 4.59).

3.1.3. Encoding of low imagery words Encoding of low imagery word pairs compared to reference task (Table 2): The prefrontal cortex (BA9/ BA10) was bilaterally activated (right: Z =4.37, left: Z = 3.72). The left cingulate gyrus (BA 31) was active (Z = 4.08) as well as the right precuneus (BA 7; Z = 4.02). Retrie6al of low imagery word pairs compared to reference task: During retrieval of abstract word pairs (Fig. 1b) significant increases in rCBF were observed in the left precuneus (BA 7: Z = 5.76). There was a significant activation (Table 4) of the bilateral prefrontal cortex (BA 10/11: right, Z =6.69; BA 10: left, Z = 5.71) and of the left cingulate gyrus (BA 29: Z =4.13). The cogniti6e subtraction ‘foreign language minus na-

3.2.2. Retrie6al of high imagery word pairs Retrie6al of high imagery word pairs compared to reference task: As shown in Fig. 2a, significant increases in regional cerebral blood flow (rCBF) were seen bilaterally in precuneus (BA 7: right Z= 3.94, left Z=4.84) and in the prefrontal cortex (BA 10/11: right Z=5.65, BA 10: left Z= 4.75). In addition, the retrieval of native words (Table 3) activated the left cingulate gyrus (BA 25, Z= 5.50) and the left cerebellar hemisphere (Z=4.93).

Table 4 Retrieval of words with low imagery content compared to reference condition Talairach coordinates

Z

BA

Region

x

y

z

Foreign language 18 −22 −2 −2

52 48 −44 −78

−4 0 20 48

6.69 5.71 4.13 5.76

10/11 10 29 7

Prefrontal cortex (right) Prefrontal cortex (left) Cingulate gyrus (left) Precuneus (left)

Mother tongue 20 0 2

56 −80 −54

−4 44 −8

4.95 5.16 4.99

10 7

Prefrontal cortex (right) Precuneus (bilateral) Cerebellum (right)

Table 5 Retrieval of word pairs as studied with cognitive subtraction Talairach coordinates

BA

Region

Retrie6al of high imagery word pairs ( foreign language — mother tongue) 50 −16 −28 3.76

20

Lower medial temporal cortex (right)

Retrie6al of low imagery word pairs ( foreign language — mother tongue) −30 24 56 5.41

8

Retrie6al of high imagery word pairs (mother tongue — foreign language) −54 4 12 4.50 4 −52 72 3.84

45 7

Broca’s area Precuneus (right)

Retrie6al of low imagery word pairs (mother tongue — foreign language) −54 8 8 4.42

44

Broca’s area

x

Z Y

z

Medial frontal gyrus (left)

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

The cogniti6e subtraction ‘mother tongue minus foreign language’ showed (Table 5 and Fig. 4a) that there was activation of Broca’s area (BA 45: left activation Z= 4.50) and in the precuneus on the right side (BA 7: Z =3.84).

53

of the two languages share common components though differential activations were found in Broca’s area and in the cerebellum as well as in the angular/ supramarginal gyri according to the language used.

4.1. Dorsolateral prefrontal cortex 3.2.3. Encoding of low imagery word pairs Encoding of low imagery word pairs as compared with the reference task (Table 2): The prefrontal cortex (BA 10/47) was bilaterally activated (right: Z =3.24, left: Z = 3.50). The cerebellar vermis was active (Z =3.53). 3.2.4. Retrie6al of low imagery word pairs Retrie6al of low imagery word pairs compared to reference task: During the retrieval of abstract word pairs (Fig. 2b), significant increases in rCBF were observed in the precuneus (BA 7: bilateral activation, Z = 5.16). Table 4 shows that there was a significant activation of the right prefrontal cortex (BA 10: Z= 4.95) and of the right cerebellar hemisphere(Z =4.99). The cogniti6e subtraction ‘mother tongue minus foreign language’ showed (Table 5 and Fig. 4b) activation of Broca’s area (BA 44: left activation Z = 4.42).

4. Discussion Our encoding results confirm previous findings which emphasize the role of the prefrontal cortex and the hippocampal formation [1,29,32,38,39,75]. The pivotal role of hippocampus in memory processes has been extensively documented [13]45[72]. There is a certain amount of overlap of the encoding network with that for retrieval in the prefrontal areas and the precuneus. These findings are in agreement with our earlier studies which showed strong functional linkages involving the prefrontal areas and the precuneus in both encoding and retrieval of word pair associations using one’s mother tongue [38]. Our study showed common cortical structures in the retrieval of Finnish and English word pairs and underlined the important role of the precuneus as a module subserving verbal memory and retrieval independent of the language used and the imagery content of the presented material. Furthermore, the prefrontal cortex was consistently active during the bilingual retrieval of word pairs under all conditions. Our results support the hypothesis of Fabbro [14] which suggests that the cortical representations of different languages in bi/multilingual subjects partly overlap. This does not exclude the possibility that within the same cortical areas distinct neural circuits independently subserve different language sets. Overlapping mechanisms are also in agreement with the results obtained with electro-cortical stimulation during neurosurgical operations [53,54]. Another crucial finding was that the brain mechanisms

Our findings of prefrontal activation during retrieval of declarative verbal information is in agreement with existing data from animals and humans [4,22,24,26,29, 75,77,78] and with neuropsychological studies of brain damaged patients [47,48,57,71,79]. The common denominator of these studies is the pivotal role of prefrontal cortex in declarative memory. Our results showed that bilateral activity patterns are observed during retrieval phases with predominance for activity in the non-dominant hemispheres in the prefrontal structures. The present findings can only be partly accounted for by the model of hemispheric encoding/retrieval asymmetry (HERA) according to which episodic retrieval processes involve right frontal regions [16,17,21,28,32,33,61,62,73,75,77,78], whereas the left frontal areas are relevant for encoding. Thus, the HERA model suggests a hemispheric encoding/retrieval asymmetry based on blood flow analyzes whereby the left prefrontal cortex is regarded to be important in retrieval of information from semantic memory and encoding of novel aspects of the retrieved information into semantic memory whereas the right prefrontal cortex is considered to be more involved in episodic memory retrieval. A significant finding of the present investigation is that bilateral activity patterns are observed during retrieval of both native and foreign word pairs. The present findings are in accordance with recent studies which suggest that the original HERA model underestimated the role of the left prefrontal cortex during memory retrieval [5,29,42]. Our findings are in line with the view that interhemispheric interactions play a crucial role during the retrieval of both native and foreign word pairs and point to the necessity for a revision of the retrieval-related aspects of the HERA model. Recently, there have been studies concerning category-related premotor and prefrontal correlates of lexical retrieval [9,27,44]. As the prefrontal processing of bilingual word pairs is concerned, our study did not show a significant difference between the categories ‘high imagery content’ and ‘low imagery content’. This applies both to the mother tongue and the foreign language.

4.2. Broca’s area Broca’s area (BA 44, 45) was activated both with the retrieval of concrete and abstract words as the cognitive subtraction mother tongue minus foreign language was

54

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

carried out. This finding will emphasize the role of Broca’s area in the retrieval of native words, irrespective of the fact whether they are concrete or abstract. In the present study, no activation of Broca’s area was found as foreign words were retrieved from memory. As for the mother tongue, our findings support the results obtained by Kim et al. [34] who showed with fMRI that Broca’s area is activated with the native language. On the other hand, there are also differences: Kim et al. [34] showed that within the frontal-lobe language-sensitive regions (Broca’s area) foreign languages acquired in adulthood are spatially separated from native languages. The difference between our results and those of Kim et al. [34] may be reconciled in part by the higher spatial resolution of the fMRI technique as compared to PET. In addition, using SPM analysis the results of several subjects were combined and averaged. This further reduces the effective resolution since individual variations in the localization of the language areas were not taken adequately into account. Another crucial difference was that the paradigm used by Kim et al. [34] was based on internal speech describing events that occurred during a specified period of the previous day. In contrast, in the present study, the retrieval of word pairs without syntax was investigated. One may argue that distinct activations within Broca’s area for native and second languages could be dependent on the use of different syntactic conditions rather than on the processing of phonetic structures of the different languages. This hypothesis is in accordance with the brain imaging study by Klein et al. [36]. The aim of their study was to investigate the neural substrates underlying phonological or semantic word generation in bilingual English– French speaking subjects. Their results suggest common neural substrates in phonological and semantic word generation tasks irrespective of whether subjects used their mother tongue or a foreign language. The subtraction analysis of the two languages revealed that in Finnish the cerebellum and Broca’s area were more activated than in English. This interesting finding deserves further investigations. It remains unclear whether the linguistic/phonological characteristics of the Finnish language are more demanding or whether the native language in general uses more sophisticated networks for word production. Further experiments are needed to analyze the neurolinguistical aspects of these languages under more complex experimental conditions using whole sentence constructions.

4.3. Precuneus This is the first study to show a consistent precuneus activation during memory retrieval of word pairs in different languages and for both highly imageable and abstract words. Our results indicate a significant and

predominantly bilateral activation of the precuneus under all experimental conditions using different languages and irrespective of the imagery content of the word pair associates. The results are in agreement with earlier brain imaging studies which showed a precuneus activation during verbal memory retrieval using the native language [16,17,29,39,49,73,74]. The precuneus activation occured for words with high and low imagery content and during visual and auditory presentation modalities [39]. Our results extend the existing evidence and emphasize the role of precuneus for both foreign and native word pairs. In spite of extensive information about lesions of the parietal lobe in various neurological disorders, particularly stroke, there are, so far, relatively few data about specific lesions of the precuneus [6,69]. However, there is anatomical evidence indicating that precuneus has connections with prefrontal [63,25] and with temporal, occipital and thalamic areas [2,65]. This connectivity creates the basis for the functional concerted action of precuneus as a multimodal association area during retrieval.

4.4. Angular/supramarginal gyri, cingulate areas Angular and supramarginal gyri are activated only when foreign words with high imagery content are retrieved. It has been shown [66] that switching the input language results in the activation of the supramarginal gyri, which may play an important role in the cognitive control of language processes. Using fMRI Yetkin et al. [82] showed that the number of activated pixels was greatest for the language in which the subject was least fluent. When native concrete words or foreign abstract words are retrieved left cingulate areas (BA 25 for native concrete words and BA 29 for foreign abstract words) are activated. The functional significance of these findings is open and further data are needed. It has recently been shown that the anterior cingulate areas in the vicinity of these areas participate in the online monitoring of performance and error detection [7,52].

4.5. Cerebellum There is significant cerebellar activation only concerning the retrieval of concrete and abstract native word pairs, whereas no cerebellar activation could be seen as foreign words were retrieved. This may refer to overlearnt automatic motor patterns associated with the mother tongue. The present results are in agreement with our earlier findings which indicated cerebellar activation during the retrieval of visually presented word pair associates [29]. Clinical lesion studies support the view of a critical role of the cerebellum in higher cognitive function [3,15].

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

55

4.6. Conclusions

Acknowledgements

Our study addressed the question whether memory processing in two languages belonging to different linguistic groups use common neural systems. We looked for shared and non-shared neural substrates in a paired-word association paradigm for such contrasting languages as Finnish and English. We reported a certain amount of overlap of the encoding neuronal network with that for retrieval in the prefrontal areas and the precuneus in both languages used. Our study showed consistent precuneus and prefrontal activation in the retrieval of Finnish and English word pairs irrespective of the imagery content of the word associates. This suggests that in the retrieval of word pairs the mother tongue and the foreign language share common prefrontal and parietal mechanisms. Our findings are in agreement with the study by Illes et al. [31] which showed that also semantic analysis in different languages (English and Spanish) is mediated by a common neural system. In the present investigation the brain areas of the two languages shared common components though differential activation patterns were found in Broca’s area, the cerebellum and the angular/supramarginal gyri. Finnish and English used in the present study represent widely different linguistic groups, non-Indo-European vs. Indo-European. We presented the word pairs in the nominative case in order to avoid the extensive morphology of the Finnish language with 14 cases for each noun [40] and to exclude unnecessary variation. Similarly, Klein et al. [37] carried out a PET study on the cerebral organization of Chinese– English bilingual subjects. They argued for shared neural substrates even for such contrasting languages as Mandarin Chinese (nonIndo-European) and English (Indo-European). There is evidence that cortical cell assemblies representing action verbs include additional areas in motor and premotor cortices [10,67]. In the present study we used neutral, non-emotional nouns; however, it is difficult to assess the emotional value of the words; different subjects may react in various ways to the same words due to their own personal life-time memories [12]. Taken together, our results argue for largely overlapping cortical areas to be activated during retrieval of native and second language word pairs. The findings advance our understanding of the neural representation that underlies multiple language functions. Further studies are needed to elucidate the neuronal mechanisms of bi/multilingual language processing. A promising perspective for future bilingual research is an integrative approach using brain imaging studies with a high spatial resolution such as fMRI, combined with techniques with a high temporal resolution, such as magnetoencephalography (MEG).

We wish to thank Professor Urpo K. Rinne, Department of Neurology, University of Turku, for his kind support. The assistance of Professor Juhani Knuuti and his staff of the PET Centre at the University of Turku is also gratefully acknowledged. U.H. had the Research Award of Alexander von Humboldt Foundation/ MRC, Academy of Finland. A.L. had an Alexander von Humboldt Grant.

References [1] Blaxton TA, Bookheimer SY, Zeffiro TA, Figlozzi CM, Gaillard WD, Theodore WH. Functional mapping of human memory using PET: comparisons of conceptual and perceptual tasks. Can J Exp Psychol 1996;50:42 – 56. [2] Blum JS, Chow KL, Pribram KH. A behavioral analysis of the organization of the parieto-temporo-preoccipital cortex. J Comp Neurol 1950;93:53 – 100. [3] Bracke-Tolkmitt R, Linden A, Canavan AGM, Rockstroh B, Scholz E, Wessel K, Diener H-C. The cerebellum contributes to mental skills. Behav Neurosci 1989;103:442 – 6. [4] Buckner RL, Petersen SE, Ojemann JG, Miezin FM, Squire LR, Raichle ME. Functional anatomical studies of explicit and implicit memory retrieval tasks. J Neurosci 1995;15:12 – 29. [5] Buckner RL. Beyond HERA: contributions of specific prefrontal brain areas to long-term memory retrieval. Psychon Bul Rev 1996;3:149 – 58. [6] Cabeza R, Mangels J, Nyberg L, Habib R, Houle S, McIntosh AR, Tulving E. Brain regions differentially involved in remembering what and when: a PET study. Neuron 1997;19:863 –70. [7] Carter CS, Braver TS, Barch DM, Botvinick MM, Noll D, Cohen JD. Anterior cingulate cortex, error detection, and the online monitoring of performance. Science 1998;280:747 –9. [8] Clark JC, Crouzel C, Meyer GJ, Strijckmans K. Current methodology for oxygen-15 production for clinical use. Appl Radiat Isot 1987;38:597 – 600. [9] Damasio H, Grabowski TJ, Tranel D, Hichwa RD, Damasio AR. A neural basis for lexical retrieval. Nature 1996;380:499 – 505. [10] Dehaene S. Electrophysiological evidence for category-specific word processing in the normal human brain. Neuroreport 1995;6:2153 – 7. [11] Dehaene S, Dupoux E, Mehler J, Cohen L, Paulesu E, Perani D, van de Moortele PF, Lehericy S, LeBihan D. Anatomical variability in the cortical representation of first and second language. Neuroreport 1997;8:3809 – 15. [12] Dolan RJ, Lane R, Chua P, Fletcher P. Dissociable temporal lobe activations during emotional episodic memory retrieval. Neuroimage 2000;11:203 –9. [13] Eldridge LL, Knowlton BJ, Furmanski CS, Bookheimer SY, Engel SA. Remembering episodes: a selective role for the hippocampus during retrieval. Nat Neurosci 2000;3:1149 –52. [14] Fabbro F, editor. The Neurolinguistics of Bilingualism. An introduction. Hove, East Sussex, UK: Psychology Press, Ltd., 1999. [15] Fiez JA, Petersen SE, Cheney MK, Raichle ME. Impaired non-motor learning and error detection associated with cerebellar damage. Brain 1992;115:155 – 78. [16] Fletcher PC, Dolan RJ, Frith CD. The functional anatomy of memory. Experientia 1995a;51:1197 – 207.

56

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57

[17] Fletcher PC, Frith CD, Grasby PM, Shallice T, Frackowiak RSJ, Dolan RJ. Brain systems for encoding and retrieval of auditory verbal memory: an in vivo study in humans. Brain 1995b;118:401 –16. [18] Friston KJ, Worsley KJ, Frackowiak RSJ, Mazziotta JC, Evans AC. Assessing the significance of focal activations using their spatial extent. Hum Brain Mapp 1994;1:210 –20. [19] Friston KJ, Ashburner J, Frith CD, Poline J-B, Heather JD, Frackowiak RSJ. Spatial registration and normalization of images. Hum Brain Mapp 1995a;2:165 –89. [20] Friston KJ, Holmes AP, Worsley KJ, Poline J-B, Frith CD, Frackowiak RSJ. Statistical parametric mapping in functional imaging: a general linear approach. Hum Brain Mapp 1995b;2:189 – 210. [21] Frith CD, Friston KJ, Liddle PF, Frackowiak RSJ. A PET study of word finding. Neuropsychologia 1991;29:1137 – 48. [22] Fuster JM, editor. Memory in the Cerebral Cortex. Cambridge (MA): MIT Press, 1995. [23] Gabrieli JDE, Brewer JB, Desmond JE, Glover GH. Separate neural bases of two fundamental memory processes in the human medial temporal lobe. Science 1997;276:264 – 6. [24] Goldman-Rakic PS. Circuitry of primate prefrontal cortex and regulation of behaviour by representational memory, In: Mountcastle VB, Plum F, editors. Handbook of Physiology, Sect I, vol. 5, P 1. Bethesda (MD): American Physiological Society, 1987. pp. 373 – 417. [25] Goldman-Rakic PS. Topography of cognition: parallel distributed networks in primate association cortex. Annu Rev Neurosci 1988;11:137 – 56. [26] Goldman-Rakic PS. The anatomy of dopamine in monkey and human prefrontal cortex. J Neural Transm Suppl 1992;36:163 – 77. [27] Grabowski TJ, Damasio H, Damasio AR. Premotor and prefrontal correlates of category-related lexical retrieval. Neuroimage 1998;7:232 – 43. [28] Grasby PM, Frith CD, Friston KJ, Bench C, Frackowiak RS, Dolan RJ. Functional mapping of brain areas implicated in auditory-verbal memory function. Brain 1993;166:1 –20. [29] Halsband U, Krause BJ, Schmidt D, Herzog H, Tellmann L, Mu¨ ller-Ga¨ rtner H-W. Encoding and retrieval in declarative learning: a positron emission tomography study. Behav Brain Res 1998;97:69 – 78. [30] Holm S, Law I, Paulson O. 3D PET activation studies with H215O bolus injection. Count rate performance and dose optimization. In: Jones T, Myers R, Bailey D, editors. Quantification of Brain Function using PET. San Diego: Academic Press, 1995. [31] Illes J, Francis WS, Desmond JE, Gabrieli JD, Glover GH, Poldrak R, Lee CJ, Wagner AD. Convergent cortical representation of semantic processing in bilinguals. Brain Lang 1999;70:347 – 63. [32] Kapur S, Craik FIM, Tulving E, Wilson A, Houle S. Neuroanatomical correlates of encoding in episodic memory: levels of processing effort. Proc Natl Acad Sci USA 1994;91:2008 – 11. [33] Kapur S, Craik FIM, Jones C, Brown GM, Houle S, Tulving E. Functional role of the prefrontal cortex in retrieval of memories: a PET study. Neuroreport 1995;6:1880 –4. [34] Kim KH, Relkin NR, Lee KM, Hirsch J. Distinct cortical areas associated with native and second languages. Nature 1997;388:171 – 4. [35] Klein D, Zatorre RJ, Milner B, Meyer E, Evans AC. Left putaminal activation when speaking a second language: evidence from PET. Neuroreport 1994;5:2295 –7. [36] Klein D, Milner B, Zatorre RJ, Meyer E, Evans AC. The neural substrates underlying word generation: a bilingual functionalimaging study. Proc Natl Acad Sci USA 1995;92:2899 – 903. [37] Klein D, Milner B, Zatorre RJ, Zhao V, Nikelski J. Cerebral organization in bilinguals: a PET study of Chinese –English verb generation. Neuroreport 1999;10:2841 –6.

[38] Krause BJ, Horwitz B, Taylor JG, Schmidt D, Mottaghy FM, Herzog H, Halsband U, Mu¨ ller-Ga¨ rtner H-W. Network analysis in episodic encoding and retrieval of word-pair associates: a PET study. Eur J Neurosci 1999a;11:3293 – 301. [39] Krause BJ, Schmidt D, Mottaghy FM, Taylor J, Halsband U, Herzog H, Tellmann L, Mu¨ ller-Ga¨ rtner H-W. Episodic retrieval activates the precuneus irrespective of the imagery content of word pair associates. A positron emission study. Brain 1999b;122:255 – 63. [40] Laine M, Niemi J, Koivuselka¨ -Sallinen P, Ahlsen E, Hyo¨ na¨ J. A neurolinguistic analysis of morphological deficits in a Finnish – Swedish bilingual aphasic. Clin Linguistics Phonetics 1994;8: 177 – 200. [41] Laine M, Rinne JO, Krause BJ, Tera¨ s M, Sipila¨ H. Left hemisphere activation during processing of morphologically complex word forms in adults. Neurosci Lett 1999;271:85 – 8. [42] Lepage M, Ghaffar O, Nyberg L, Tulving E. Prefrontal cortex and episodic memory retrieval mode. Proc Natl Acad Sci USA 2000;97:506 – 11. [43] Lewellen TK, Kohlmyer SG, Miyaoka RS, Kaplan MS, Stearns CW, Schubert SF. Investigation of the Performance of the General Electric ADVANCE Positron Emission Tomograph in 3D Mode. IEEE Trans Nucl Sci 1996;43:2199 – 206. [44] Martin A, Wiggs CL, Ungerleider LG, Haxby JV. Neural correlates of category-specific knowledge. Nature 1996;379:649 –52. [45] Mellet E, Tzourio-Mazoyer N, Bricogne S, Mazoyer B, Kosslyn SM, Denis M. Functional anatomy of high-resolution visual imagery. J Cogn Neurosci 2000;12:98 – 109. [46] Mesulam MM. Large-scale neurocognitive networks and distributed processing for attention, language, and memory. Ann Neurol 1990;28:597 – 613 Review. [47] Milner B, Petrides M, Schmidt ML. Frontal lobes and the temporal organization of memory. Hum Neurobiol 1985;4:137 – 42. [48] Moscovitch M. Memory and working-with-memory: a component process model based on modules and central systems. J Cogn Neurosci 1992;4:257 – 67. [49] Mottaghy FM, Shah NJ, Krause BJ, Schmidt D, Halsband U, Ja¨ ncke L, Mu¨ ller-Ga¨ rtner H-W. Neuronal correlates of encoding and retrieval in episodic memory during a paired-word association learning task: a functional magnetic resonance imaging study. Exp Brain Res 1999;128:332 – 42. [50] Neville HJ, Coffey SA, Lawson DS, Fischer A, Emmorey K, Bellugi U. Neural systems mediating American sign language: effects of sensory experiences and age of acquisition. Brain Lang 1997;57:285 – 308. [51] Neville HJ, Mills DL, Lawson DS. Fractionating language: different neural subsystems with different sensitive periods. Cereb Cortex 1992;2:244 – 58. [52] Ochsner KN, Kosslyn SM, Rees Cosgrove G, Cassem EH, Price BH, Nierenberg AA, Rauch SL. Deficits in visual cognition and attention following bilateral anterior cingulotomy. Neuropsychologia 2001;39:219 – 30. [53] Ojemann GA, Whitaker HA. The bilingual brain. Arch Neurol 1978;35:409 – 12. [54] Ojemann GA. Cortical organization of language. J Neurosci 1991;11:2281 – 7 August. [55] Paivio A, Yuille JC, Smythe PC. Stimulus and response abstractness, imagery and meaningfulness and reported mediators in paired-associate learning. Can J Psychol 1966;20:362 – 77. [56] Paivio A, Yuille JC, Madigan SA. Concreteness, imagery and meaningfulness values for 925 nouns. J Exp Psychol 1968;76(Suppl. 1):1 – 25. [57] Paradis M. Neurolinguistic aspects of implicit and explicit memory: implications for bilingualism and second language acquisition. In: Ellis N, editor. Implicit and Explicit Language Learning. London: Academic Press, 1994:393 – 419.

U. Halsband et al. / Beha6ioural Brain Research 132 (2002) 47–57 [58] Paradis M. Generalizable outcomes of bilingual aphasia research. Folia Phoniatr Logop 2000;52:54 –64. [59] Perani D, Dehaene S, Grassi F, Cohen L, Cappa SF, Dupoux E, Fazio F, Mehler J. Brain processing of native and foreign languages. Neuroreport 1996;7:2439 – 44. [60] Perani D, Paulesu E, Galles NS, Dupoux E, Dehaene S, Bettinardi V, Cappa SF, Fazio F, Mehler J. The bilingual brain. Proficiency and age of acquisition of the second language. Brain 1998;121:1841 – 52. [61] Petersen SE, Fox PT, Posner MI, Mintun M, Raichle ME. Positron emission tomographic studies of the cortical anatomy of single-word processing. Nature 1988;331:585 –9. [62] Petersen SE, Fox PT, Snyder AZ, Raichle ME. Activation of extrastriate and frontal cortical areas by visual words and wordlike stimuli. Science 1990;249:1041 –4. [63] Petrides M, Pandya DN. Projections to the frontal cortex from the posterior parietal region in the rhesus monkey. J Comp Neurol 1984;228:105 – 16. [64] Pouratian N, Bookheimer SY, Oı´Farrell A, Sicotte NL, Cannestra AF, Becker D, Toga AW. Optical imaging of bilingual cortical representations. J Neurosurg 2000;93:676 –81. [65] Pribram HB, Barry J. Further behavioral analysis of parieto – temporo-preoccipital cortex. J Neurophysiol 1956;19:99 – 106. [66] Price CJ, Green DW, von Studnitz R. A functional imaging study of translation and language switching. Brain 1999;122:2221– 35. [67] Pulvermu¨ ller F, Lutzenberger W, Preissl H. Nouns and verbs in the intact brain; evidence from event-related potentials and high-frequency cortical responses. Cereb Cortex 1999;9:497 – 506. [68] Roberts PM, Le Dorze G. Bilingual aphasia: semantic organization, strategy use, and productivity in semantic verbal fluency. Brain Lang 1998;65:287 – 312. [69] Ross ED. Left medial parietal lobe and receptive language functions: mixed transcortical aphasia after left anterior cerebral artery infarction. Neurology 1980;30:144 –51. [70] Saukkonen P, Haipus M, Niemikorpi A, Sulkala H, editors. Suomen Kielen Taajuussanasto (Frequency Dictionary of Finnish). Porvoo: WSOY, 1979.

57

[71] Schacter DL. Memory, amnesia and frontal lobe dysfunction. Psychobiology 1987;15:21 – 36. [72] Schacter DL, Curran T, Reiman EM, Chen K, Bandy DJ, Frost JT. Medial temporal lobe activation during episodic encoding and retrieval: a PET study. Hippocampus 1999;9:575 – 81. [73] Shallice T, Fletcher P, Frith CD, Grasby P, Frackowiak RSJ, Dolan RJ. Brain regions associated with acquisition and retrieval of verbal episodic memory. Nature 1994;368:633 – 5. [74] Schmidt D, Krause BJ, Mottaghy FM, Halsband U, Herzog H, Tellman L, Mu¨ ller-Ga¨ rtner H-W. Brain systems engaged in encoding and retrieval of word pair associates independent of their imagery content or presentation modalities. Neuropsychologia 2001, in press. [75] Squire LR, Ojeman JG, Miezin FM, Petersen SE, Videen TO, Raichle ME. Activation of the hippocampus in normal humans: a functional anatomical study of memory. Proc Natl Acad Sci USA 1992;89:1837 – 41. [76] Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain. In: 3-Dimensional Proportional System: An Approach to Cerebral Imaging. Stuttgart: Thieme, 1988. [77] Tulving E, Kapur S, Markowitsch HJ, Craik FIM, Habib R, Houle S. Neuroanatomical correlates of retrieval in episodic memory: auditory sentence recognition. Proc Natl Acad Sci USA 1994a;91:2012 – 5. [78] Tulving E, Kapur S, Craik FIM, Markowitsch HJ, Houle S. Hemispheric encoding/retrieval asymmetry in episodic memory: positron emission tomography findings. Proc Natl Acad Sci USA 1994b;91:2016 – 20. [79] Warrington EK, Weiskrantz L. Amnesia: a disconnection syndrome? Neuropsychologia 1982;20:233 – 49. [80] Weber-Fox CM, Neville HJ. Maturational constraints on functional specializations for language processing: ERP and behavioral evidence in bilingual speakers. J Cogn Neurosci 1996;8:231 – 56. [81] Wechsler D. A standardized memory scale for clinical use. J Psychol 1945;19:87 – 95. [82] Yetkin O, Zerrin Yetkin F, Haughton VM, Cox RW. Use of functional MR to map language in multilingual volunteers. Am J Neuroradiol 1996;17:473 – 7.