Using Digital Twins to Predict Recovery and Decline in Neurological Disorders

In this viewpoint, we describe how digital twins, individually configured computational models of a person's unique language learning history and neurological status, can be used to predict recovery and decline in neurological disorders, with a particular focus on their potential for neurorehabilitation. We first situate digital twins within the broader evolution of computational modeling research in cognitive science and neurorehabilitation. We then introduce BiLex, a digital twin framework that uses self-organizing maps (SOMs) to model bilingual language processing through interconnected semantic and phonetic representations for two languages, and we explain why bilingualism serves as both a rigorous test case for the methodology and a clinically important application. We then describe the neurorehabilitation applications of BiLex including modeling stroke and aphasia that include individual impairment profiles, recovery, treatment-response prediction, and selection of the optimal treatment language, as well as simulating progressive decline in dementia. We summarize that the framework offers theory-driven model construction to improve our understanding of neurorehabilitation. Digital twins represent a paradigm shift enabling individualized assessment, treatment planning, and progress monitoring in neurological rehabilitation.