English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The difference between the chronological and biological brain age of asubject can be an important biomarker for neurodegenerative diseases, thusbrain age estimation can be crucial in clinical settings. One way toincorporate multimodal information into this estimation is through populationgraphs, which combine various types of imaging data and capture theassociations among individuals within a population. In medical imaging,population graphs have demonstrated promising results, mostly forclassification tasks. In most cases, the graph structure is pre-defined andremains static during training. However, extracting population graphs is anon-trivial task and can significantly impact the performance of Graph NeuralNetworks (GNNs), which are sensitive to the graph structure. In this work, wehighlight the importance of a meaningful graph construction and experiment withdifferent population-graph construction methods and their effect on GNNperformance on brain age estimation. We use the homophily metric and graphvisualizations to gain valuable quantitative and qualitative insights on theextracted graph structures. For the experimental evaluation, we leverage the UKBiobank dataset, which offers many imaging and non-imaging phenotypes. Ourresults indicate that architectures highly sensitive to the graph structure,such as Graph Convolutional Network (GCN) and Graph Attention Network (GAT),struggle with low homophily graphs, while other architectures, such asGraphSage and Chebyshev, are more robust across different homophily ratios. Weconclude that static graph construction approaches are potentially insufficientfor the task of brain age estimation and make recommendations for alternativeresearch directions.

A Comparative Study of Population-Graph Construction Methods and Graph Neural Networks for Brain Age Regression