LU Hanna, PhD | The Chinese University of Hong Kong; LI Yu, PhD; ZHANG Li, PhD
Competition Sponsor: Research Grants Council of the Hong Kong Special Administrative Region, China
Award Year: 2023
Towards healthy longevity and successful ageing, brain age, rather than chronological age, has been increasingly used as a phenotype to evaluate the brain fitness and predict the cognitive maintenance at individual level. Brain age is closely related to a bunch of factors, including brain structure, functional connectivity, and cerebrovascular health. Recently, several models of brain age have been identified and strongly related to neurodegeneration and its progression. However, the utility of these models may be blunted by their reliance on the defined features drawn from single brain feature (e.g., gray matter). The current limitations lead to the motivation for developing an integrative model to evaluate brain age. In this project, we propose a novel deep learning framework capable of exploiting network-based radiomic features to learn the relationships between brain structure and function across the adult lifespan such that more accurate brain age could be predicted. This method will be trained on two large-scale cohort datasets and then tested on a longitudinal dataset of Hong Kong SuperAgers, utilizing novel brain age model to validate and improve the prediction accuracy. Multi-modal MRI data will be analyzed based on semi-automated platform, generated individualized radiomic models and then achieve the prediction of brain age. This project can interpret the individual differences between chronological age, brain age and cognitive maintenance for senior adults. Furthermore, it can be used to evaluate the cognitive maintenance, predict the conversion from healthy ageing to dementia and disease progression, serve as a personalized biomarker for detecting, monitoring and preventing neurodegenerative diseases.