Identificare le fasi prodromiche della demenza attraverso l’analisi computazionale del linguaggio: quindici anni di ricerche sull’italiano

Early identification of cognitive impairment, especially in the subclinical phases of neurodegenerative diseases, is a key challenge for modern healthcare. Traditional screening tools often fail to detect the subtle language and cognitive changes that precede dementia. Given the strong evidence that language is one of the earliest faculties affected by neurodegeneration, linguistic analysis is emerging as a promising tool for early diagnosis. This paper presents a novel approach developed by the University of Bologna’s Department of Classical Philology and Italian Studies, which uses Natural Language Processing (NLP) and Artificial Intelligence (AI) to automatically analyze patients’ speech. By extracting linguistic biomarkers from semi-spontaneous vocal recordings, the method aims to support early detection of cognitive decline and offers a preliminary communicative profile of Italian-speaking individuals with dementia.