Predicting language treatment response in bilingual aphasia using neural network-based patient models

摘要: Predicting language therapy outcomes in bilinguals with aphasia (BWA) remains challenging due to the multiple pre- and poststroke factors that determine deficits recovery of their two languages. Computational models simulate impairment treatment BWA can help predict response identify optimal for treatment. Here we used BiLex computational model behavioral profile a retrospective sample 13 Spanish-English who received one Specifically, simulated prestroke naming ability each language, treated untreated language. predicted effects accurately robustly captured different degrees cross-language generalization BWA. Our cross-validation approach further demonstrated generalizes patients whose data were not training. These findings support potential suggest modeling may be helpful guide individually tailored rehabilitation plans this population.