Several works show a relationship between sensory loss and cognitive performance and between sensory loss and dementia. The DEMETRA project aims to provide a better understanding of the relationship between sensory loss and cognitive loss in elderly individuals with dementia or at risk of dementia, using objective technologies to measure sensory deficits. The researcher's activities will concern the development of eXplainable artificial intelligence algorithms (XAI) to predict different types of dementia (Alzheimer's disease, fronto-temporal dementia and vascular dementia) at different stages (mild cognitive impairment and no cognitive impairment) from several clinical information. Furthermore, the researcher will develop a connectomic map of morphological features describing pattern differences compared to controls with no cognitive impairment using sophisticated deep learning algorithms.