Laterality and language experience

Abstract

A meta-analysis was conducted on studies that examined hemispheric functional asymmetry for language in brain-intact monolingual and bilingual adults. Data from 23 laterality studies that directly compared bilingual and monolingual speakers on the same language were analysed (n = 1234). Variables examined were language experience (monolingual, bilingual), experimental paradigm (dichotic listening, visual hemifield presentation, and dual task) and, among bilinguals, the influence of second language proficiency (proficient vs nonproficient) and onset of bilingualism (early, or before age 6; and late, or after age 6). Overall, monolinguals and late bilinguals showed reliable left hemisphere dominance, while early bilinguals showed reliable bilateral hemispheric involvement. Within bilinguals, there was no reliable effect of language proficiency when age of L2 acquisition was controlled. The findings indicate that early learning of one vs. two languages predicts divergent patterns of cerebral language lateralisation in adulthood.

Portions of this research were presented at the 4th International Symposium on Bilingualism held in Tempe, AZ, 2003.

Portions of this research were presented at the 4th International Symposium on Bilingualism held in Tempe, AZ, 2003.

Acknowledgments

We thank Colin Allen, Winfred Arthur, Heather Bortfeld, Wendy Francis, Michel Paradis, Loraine Obler, Steven M. Smith, Daniel Voyer, and Wendy Wood for constructive comments and/or technical assistance.

Notes

Portions of this research were presented at the 4th International Symposium on Bilingualism held in Tempe, AZ, 2003.

¹Data from neuroimaging and electrophysiological studies were excluded from the meta-analysis because of the inappropriateness of including studies that are incompatible in terms of statistical, methodological, and conceptual grounds (see Lipsey & Wilson, 2001).

²Web of Science version 4.1 forward citation search; Science Citation Index (Expanded) (1982–present) forward citation search; Social Sciences Citation Index (1982–present) forward citation search; Arts and Humanities Citation Index (1982–present) forward citation search; Biology Digest (1989–present); Biological Sciences (1982–present); Digital Dissertations (1861–present); ERIC; Linguistic and Language Behavioral Abstracts (1973–present); MedLine (1989–present); OVID; PsycINFO; Web Resources Related to the Social Sciences; Alternative Press Index of NISC and Biblioline; Dissertation Abstracts; Sociological Abstracts (1963–present); NTIS; ASFA1: Biological Sciences and Living Resources.

³ Brain & Language; Brain & Cognition; Memory & Cognition; Journal of Experimental Psychology: Learning, Memory and Cognition; NeuroImage; Psychological Bulletin & Review; Journal of Memory & Language.

⁴Caution is advised when interpreting results based on fewer than five data points, and some argue that at least 10 data points are needed (see Arthur, Bennet, & Huffcutt, 2001; Lipsey & Wilson, 2001).

⁵An alternative would have been to use a random effects model, which assumes an inherent difference between studies. Given that the sample consisted completely of variations on studies designed specifically to test hemispheric involvement, and the assumption that the bulk of the variance could be explained by moderating variables, the random effects model was rejected as an option.

⁶In fixed effects categorical models, it is appropriate to make conclusions concerning the divisions of the moderating variables that have been coded from the included studies (cf. Wood & Quinn, 2003).

⁷Following Cohen (1992), ds of 0.20, 0.30, and 0.50 represent small, medium, and large effect sizes, respectively.