Principal Investigator: Dr Nambi Nallasamy
Department: University of MichiganTags: 59418, biological age, Dementia, Fundus photography, lifestyle, retinal OCT
It is well known that people age differently. Promoting healthy longevity requires that we learn from those who age well, and that we slow the aging process in those who seemingly age too quickly. However, to do so it is necessary to be able to understand which patients are at risk of aging poorly (through the development of dementia or through premature death). The human retina provides a unique window into the health of a human being, as it allows us to directly view components of the central nervous system and cardiovascular system at the same time through non-invasive photographs.
Deep learning is a powerful tool that uses the predictive power of all features contained in an image, and is not restricted to those characteristics that researchers can see and presume to be relevant. Prior research has found that deep learning can be applied to retinal images to identify many non-ocular health traits. In this study, using separate training and validation sets of retinal images from individuals in the UK Biobank, we will develop and test deep learning algorithms to predict risk of dementia and premature death.
This research may catalyze critical advances in biogerontology and medicine. For example, accurate identification of risk of dementia and premature death may be used to measure the efficacy of interventions that aim to slow the aging process; to identify resilience factors in those who age slowly; and to improve scientific understanding of the cellular and molecular changes that drive physiologic aging. Consequently, this project holds the potential to transform the study of healthy longevity.