TY - CHAP
T1 - A Comparative Study of Population-Graph Construction Methods and Graph Neural Networks for Brain Age Regression
AU - Bintsi, Kyriaki Margarita
AU - Mueller, Tamara T.
AU - Starck, Sophie
AU - Baltatzis, Vasileios
AU - Hammers, Alexander
AU - Rueckert, Daniel
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - The difference between the chronological and biological brain age of a subject can be an important biomarker for neurodegenerative diseases, thus brain age estimation can be crucial in clinical settings. One way to incorporate multimodal information into this estimation is through population graphs, which combine various types of imaging data and capture the associations among individuals within a population. In medical imaging, population graphs have demonstrated promising results, mostly for classification tasks. In most cases, the graph structure is pre-defined and remains static during training. However, extracting population graphs is a non-trivial task and can significantly impact the performance of Graph Neural Networks (GNNs), which are sensitive to the graph structure. In this work, we highlight the importance of a meaningful graph construction and experiment with different population-graph construction methods and their effect on GNN performance on brain age estimation. We use the homophily metric and graph visualizations to gain valuable quantitative and qualitative insights on the extracted graph structures. For the experimental evaluation, we leverage the UK Biobank dataset, which offers many imaging and non-imaging phenotypes. Our results 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 as GraphSage and Chebyshev, are more robust across different homophily ratios. We conclude that static graph construction approaches are potentially insufficient for the task of brain age estimation and make recommendations for alternative research directions.
AB - The difference between the chronological and biological brain age of a subject can be an important biomarker for neurodegenerative diseases, thus brain age estimation can be crucial in clinical settings. One way to incorporate multimodal information into this estimation is through population graphs, which combine various types of imaging data and capture the associations among individuals within a population. In medical imaging, population graphs have demonstrated promising results, mostly for classification tasks. In most cases, the graph structure is pre-defined and remains static during training. However, extracting population graphs is a non-trivial task and can significantly impact the performance of Graph Neural Networks (GNNs), which are sensitive to the graph structure. In this work, we highlight the importance of a meaningful graph construction and experiment with different population-graph construction methods and their effect on GNN performance on brain age estimation. We use the homophily metric and graph visualizations to gain valuable quantitative and qualitative insights on the extracted graph structures. For the experimental evaluation, we leverage the UK Biobank dataset, which offers many imaging and non-imaging phenotypes. Our results 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 as GraphSage and Chebyshev, are more robust across different homophily ratios. We conclude that static graph construction approaches are potentially insufficient for the task of brain age estimation and make recommendations for alternative research directions.
KW - Brain age regression
KW - Graph Neural Networks
KW - Population graphs
UR - http://www.scopus.com/inward/record.url?scp=85188733418&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-55088-1_6
DO - 10.1007/978-3-031-55088-1_6
M3 - Conference paper
AN - SCOPUS:85188733418
SN - 9783031550874
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 64
EP - 73
BT - Graphs in Biomedical Image Analysis, and Overlapped Cell on Tissue Dataset for Histopathology - 5th MICCAI Workshop, GRAIL 2023 and 1st MICCAI Challenge, OCELOT 2023, Held in Conjunction with MICCAI 2023, Proceedings
A2 - Ahmadi, Seyed-Ahmad
A2 - Pereira, Sérgio
PB - Springer Science and Business Media Deutschland GmbH
T2 - 5th Workshop on GRaphs in biomedicAl Image anaLysis Satellite event at MICCAI, GRAIL 2023 and 1st Cell Detection from Cell-Tissue Interaction challenge in MICCAI, OCELOT 2023 Held in Conjunction with International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2023
Y2 - 8 October 2023 through 12 October 2023
ER -