OHBM 2022 | 161 | Talk | Yu Xiao | Predicting Individual Brain Regional Atrophy Progression Using …
Title: Predicting Individual Brain Regional Atrophy Progression Using Functional Connectome.
Session: Talk
Speaker: Yu Xiao
The network-based neurodegeneration hypothesis posits that neurodegeneration of Alzheimer’s disease (AD) targets specific large-scale neuronal networks, which mirrors the default mode functional network in healthy participants. Our previous work has demonstrated the potential of the brain functional connectome in predicting regional atrophy in neurodegenerative disorders. Brown et al. added further evidence at the individual level by predicting frontotemporal lobar degeneration using a generalized additive model. Yet, no individual nodal atrophy progression prediction model has been developed in AD. Little is known about the possible differential contribution of intra-network and inter-network connectivity to downstream neurodegeneration. Further, existing methods mostly rely on nodal graph theoretical measures which provide a restricted view of the functional connectome. Here, we developed an efficient and interpretable individualized, longitudinal nodal atrophy prediction model in the AD spectrum, leveraging the comprehensive graph structure and properties of the brain functional connectome. Our model significantly outperformed benchmark models (R = 0.46 ± 0.05). We identified important FC features that contributed most to network-level atrophy prediction using GNNExplainer. We found that 1) atrophy was most strongly predicted by FC of its own network, followed by hippocampus, default, and control networks; and 2) the prediction pattern of FC varied with cognitive stages in Aβ+ participants. Next, we showed that the estimated nodal atrophy from our model predicted cognitive decline (mini-mental state examination (MMSE), clinical dementia rating sum of boxes (CDR-SB)) with high accuracy ((MMSE: R = 0.68 ± 0.05, CDR-SB: R = 0.61 ± 0.09). Lastly, the prediction results were also validated in an independent Singapore dataset, suggesting the generalizability and validity of our study. In addition to providing evidence for network-based degeneration in AD, the findings represent a key step along the path toward developing prognostic assays for disease progression prediction.Write your abstract here.
Watch video OHBM 2022 | 161 | Talk | Yu Xiao | Predicting Individual Brain Regional Atrophy Progression Using … online without registration, duration hours minute second in high quality. This video was added by user Organization for Human Brain Mapping 30 July 2024, don't forget to share it with your friends and acquaintances, it has been viewed on our site 12 once and liked it 0 people.