The combination of an aging bilingual population combined with a projected increase in individuals with dementia increases the urgency to understand the interaction between bilingualism and neurodegenerative disorders. There are a cluster of neurodegenerative disorders lead to cognitive decline and dementia, such as Alzheimer?s disease  and fronto-temporal dementia spectrum disorders [2, 3], including primary progressive aphasia (PPA) [4, 5] and posterior cortical atrophy  and are all characterized by language deficits. It is therefore, important to understand mechanisms of language decline in these individuals in order to improve assessments or develop any restorative or facilitative approaches to help these individuals lead fulfilling lives. Further, given the complex multitude of factors such as the age of acquisition of the second language, the relative language exposure/proficiency in the two languages, the pathophysiology of the neurodegeneration, it is virtually impossible if not impractical to obtain a clear understanding of the nature of bilingual language decline in neurodegenerative disorders. As a novel alternative, in a previously funded project (NIH/NIDCD R21DC009446) and current project (U01DC014922), we have developed BILEX, a computational model that can accurately simulate lexical access in healthy bilinguals while accounting for individual differences in the age of acquisition of the second language, and the relative amounts of lifetime use and exposure to each language. The model also accounts for the effects of these factors on the representational organization of the bilingual mental lexicon at the individual level. In this supplement, we aim to extend this model to simulate bilingual language impairment in individuals with neurodegenerative disorders (specifically PPA and AD/MCI).
In Aim 1, we will simulate semantic decline in the BILEX model by damaging the semantic map. Performance on measures of object knowledge and lexical access in the model will be validated against human data collected from 20 bilingual individuals with semantic dementia. We expect to demonstrate that after appropriate damage to the semantic representation maps, BILEX can accurately simulate deficits of object knowledge and lexical access deficits in bilingual dementia.
In Aim 2, we will evaluate this model?s ability to predict decline in naming and object knowledge, by comparing longitudinal data (three time points) in bilingual individuals? with dementia with the model?s prediction of decline based on data from two time points. We expect to demonstrate that progressively increasing damage to the semantic representation maps will accurately simulate and predict performance decline in bilingual individuals.
The proposed work will determine the nature of semantic impairment and decline in L1 and L2 in individuals with bilingual semantic dementia, particularly in individuals with PPA and MCI. In addition, the computational model allows specification of variables such as AoA, language exposure/proficiency, the mechanism of impairment across different subtypes of dementia and even the rate of progression of the decline, giving us, for the first time, a tractable possible solution to understanding language decline in individuals with dementia.