Abstract
In this study, we investigated dual-language decline in non-balanced bilinguals with probable Alzheimer's disease (AD) both longitudinally and cross-sectionally. We examined patients' naming accuracy on the Boston Naming Test (BNT: Kaplan et al., 1983) over three testing sessions (longitudinal analysis) and compared their performance to that of matched controls (cross-sectional analysis). We found different longitudinal and cross-sectional patterns of decline: Longitudinally, the non-dominant language seemed to decline more steeply than the dominant language, but, cross-sectionally, differences between patients and controls were larger for the dominant than for the non-dominant language, especially at the initial testing session. This differential pattern of results for cross-sectional versus longitudinal decline was supported by correlations between decline measures and BNT item characteristics. Further studies will be needed to better characterize the nature of linguistic decline in bilinguals with AD; however, these results suggest that representational robustness of individual lexical representations, rather than language membership, might determine the time course of decline for naming in bilinguals with AD.