Search Papers | Poster Sessions | All Posters
Poster A156 in Poster Session A - Tuesday, August 6, 2024, 4:15 – 6:15 pm, Johnson Ice Rink
Modelling Brain Atrophy Dynamics Enhances Predicting Cognitive Decline in Alzheimer's Disease Continuum
Wooseok Jung1 (), Saehyun Kim1, Dong-Hee Kim1, Won Hwa Kim2; 1VUNO Inc., Seoul, South Korea, 2POSTECH, South Korea
Alzheimer’s disease (AD) is recognized as a continuum of cognitive decline with underlying biological changes, and predicting the disease progression is crucial. Brain atrophy status obtained from volumetric MRI is pivotal for assessing disease severity and prognosis during the continuum, but modelling its longitudinal change and its relationship with the progression has been underexplored. This study proposes a novel deep learning-based method that precisely models the atrophy dynamics across 62 cortical and subcortical regions of mild cognitive impairment (MCI) subjects collected from the ADNI database, followed by a unique training scheme to add the effects of beta-amyloid protein deposition on atrophy and to model subject-specific dynamical features. Furthermore, we directly implement the dynamics into the MCI to dementia conversion prediction task. Our findings demonstrate the feasibility of modelling atrophy dynamics using deep learning and suggest that leveraging dynamics representation (DR) enhances the conversion prediction.
Keywords: Implicit neural representation Atrophy dynamics Mild cognitive impairment Alzheimer's disease