Articulation in early and late bilinguals’ two languages: evidence from functional magnetic resonance imaging Cheryl Frenck-Mestre, 1,2,CA Jean Luc Anton, 1,3 Muriel Roth, 1,3 Jyotsna Vaid 4 and Franc° ois Viallet 5 1 Centre National de Recherche Scienti¢que, France; 2 Laboratoire Parole et Langage, Universite¤ de Provence, Aix-en-Provence, France; 3 Centre d’IRM fonctionnelle de Marseille, Marseille, France; 4 Texas A&M University, College Station, Texas, USA; 5 Service de Neurology, Ho“ pitaux du Pays d’Aix, Aix-en-Provence, France CA,1 Corresponding Author: frenck@up.univ-aix.fr Received 5 February 2005; accepted 2 March 2005 The network of cortical and subcortical regions that contribute to articulation was examined in bilinguals using functional magnetic resonance imaging. Participants were all £uent in French and Eng- lish: half were bilingual from birth and half were ‘late bilinguals’ who had learned French after the age of 12. Overt articulation resulted in the bilateral activation of the motor cortex, basal ganglia and cerebellum, and also the supplementary motor area, independent of the language spoken. Furthermore, the threshold and extent of the network involved in articulation was identical for the two bilin- gual groups with the exception of greater variation in the left puta- men for the late bilinguals. These data challenge claims that age of acquisition results in fundamental di¡erences in the neural sub- strates that subserve language in bilinguals. NeuroReport16:761^765  c 2005 Lippincott Williams & Wilkins. Key words: Articulation; Early bilinguals; Functional magnetic resonance imaging; Late bilinguals; Putamen INTRODUCTION Advances in imaging technology have allowed us to gain a far better understanding of the interactions involved between different regions of the brain when we engage in even highly automated tasks such as articulating speech. Evidence from monolinguals suggests that articulation involves a network of cortical and subcortical structures including the premotor and primary motor cortex, the supplementary motor area (SMA), the basal ganglia and the cerebellum [1,2]. The role of the anterior insula has also been suggested [3,4]. The extent to which these different regions are involved during speech varies as a function of several factors. Overt speech produces bilateral activation of the motor cortex, cerebellum and basal ganglia, whereas covert speech produces unilateral activation of the left motor cortex and comparatively reduced activation of the basal ganglia [5–7]. At slow rates of speech, the cerebellum is apparently not actively involved in a simple articulation task, whereas at rates that approximate normal speech it is [8]. When a particular rhythm is imposed upon simple sequences, the left putamen is involved in articulation, whereas for monotonic sequences (compared with a perceptual baseline) it is not systematically active [9]. Given the presence of such variability in brain activation during articulation in monolinguals, one might expect as much or more variability in the neural regions associated with articulation in bilingual speakers, particularly for those who learned a second language ‘late’ and who still bear the phonological mark of their first-learned language. Several dozen neuroimaging studies of language processing in bilinguals have been published [10]. Only a handful, however, have examined the motor aspects of speech. Klein and colleagues [11] reported a positron emission tomogra- phy (PET) study in which proficient English–French late bilinguals overtly repeated and translated words. Highly similar patterns of cortical and subcortical activation were found across the two languages with the exception of the left putamen [12], where higher activation was found in the second language than in the native language. The authors attributed this difference to greater articulatory demands in the second, later learned language. Three subsequent PET studies, however, did not support this claim. No difference in putamen activation across languages was found in Chinese–English late bilinguals on a repetition task [13], or in German–English late bilinguals on a translation task [14]. Klein et al. [15] later reported that variation in putamen activation was bound to particular word properties (asso- ciated with articulating French), rather than specifically a second language effect. These PET studies thus concur in showing overlapping areas of activation across languages during speech, with the possible exception of the left putamen. Most functional magnetic resonance imaging (fMRI) studies in bilinguals also show overlapping rather than spatially distinct regions activated during language tasks; however, these studies have not looked specifically at articulation. Where differences are found, they appear to be associated with late onset of bilingualism and/or less proficiency in the second language; however, here too there is opposing evidence [16]. Thus, currently, no clear consensus exists about whether neural activation linked to BRAIN IMAGING NEUROREPORT 0959-4965  c Lippincott Williams & Wilkins Vol 16 No 7 12 May 2005 761 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.