Back to the basics with inclusion of clinical domain knowledge - A simple, scalable and effective model of Alzheimer’s Disease classification

Sarah C. Brüningk, Felix Hensel, Louis P. Lukas, Merel Kuijs, Catherine R. Jutzeler, Bastian Rieck
Proceedings of the 6th Machine Learning for Healthcare Conference, PMLR 149:730-754, 2021.

Abstract

On high-resolution structural magnetic resonance (MR) images Alzheimer’s disease (AD) is pathologically characterised by brain atrophy and an overall loss of brain tissue connectivity. In this study, we harness such prior clinical domain knowledge to evaluate MR image-based classification of AD patients from healthy controls using deliberately simple convolutional neural network (CNN) architectures. In addition to evaluating CNN performance on high resolution structural MR imaging data, we consider topological feature representations thereof to evaluate structural connectivity. We perform an ablation study, combined with model interpretability analysis, to evaluate the relevance of the specific image region used for classification. Notably, we find that by choosing a meaningful data representation comprising the left hippocampus, we achieve competitive performance (accuracy 84 ± 7%) comparable to far more complex, heavily parameterised machine learning architectures. This implies that clinical domain knowledge may overrule the importance of model architecture design in the case of AD classification. This opens up new possibilities for interpretable architectures and simplifies model training in terms of computational cost and hardware requirements.

Cite this Paper


BibTeX
@InProceedings{pmlr-v149-bruningk21a, title = {Back to the basics with inclusion of clinical domain knowledge - A simple, scalable and effective model of Alzheimer’s Disease classification}, author = {Br\"uningk, Sarah C. and Hensel, Felix and Lukas, Louis P. and Kuijs, Merel and Jutzeler, Catherine R. and Rieck, Bastian}, booktitle = {Proceedings of the 6th Machine Learning for Healthcare Conference}, pages = {730--754}, year = {2021}, editor = {Jung, Ken and Yeung, Serena and Sendak, Mark and Sjoding, Michael and Ranganath, Rajesh}, volume = {149}, series = {Proceedings of Machine Learning Research}, month = {06--07 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v149/bruningk21a/bruningk21a.pdf}, url = {https://proceedings.mlr.press/v149/bruningk21a.html}, abstract = {On high-resolution structural magnetic resonance (MR) images Alzheimer’s disease (AD) is pathologically characterised by brain atrophy and an overall loss of brain tissue connectivity. In this study, we harness such prior clinical domain knowledge to evaluate MR image-based classification of AD patients from healthy controls using deliberately simple convolutional neural network (CNN) architectures. In addition to evaluating CNN performance on high resolution structural MR imaging data, we consider topological feature representations thereof to evaluate structural connectivity. We perform an ablation study, combined with model interpretability analysis, to evaluate the relevance of the specific image region used for classification. Notably, we find that by choosing a meaningful data representation comprising the left hippocampus, we achieve competitive performance (accuracy 84 ± 7%) comparable to far more complex, heavily parameterised machine learning architectures. This implies that clinical domain knowledge may overrule the importance of model architecture design in the case of AD classification. This opens up new possibilities for interpretable architectures and simplifies model training in terms of computational cost and hardware requirements.} }
Endnote
%0 Conference Paper %T Back to the basics with inclusion of clinical domain knowledge - A simple, scalable and effective model of Alzheimer’s Disease classification %A Sarah C. Brüningk %A Felix Hensel %A Louis P. Lukas %A Merel Kuijs %A Catherine R. Jutzeler %A Bastian Rieck %B Proceedings of the 6th Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2021 %E Ken Jung %E Serena Yeung %E Mark Sendak %E Michael Sjoding %E Rajesh Ranganath %F pmlr-v149-bruningk21a %I PMLR %P 730--754 %U https://proceedings.mlr.press/v149/bruningk21a.html %V 149 %X On high-resolution structural magnetic resonance (MR) images Alzheimer’s disease (AD) is pathologically characterised by brain atrophy and an overall loss of brain tissue connectivity. In this study, we harness such prior clinical domain knowledge to evaluate MR image-based classification of AD patients from healthy controls using deliberately simple convolutional neural network (CNN) architectures. In addition to evaluating CNN performance on high resolution structural MR imaging data, we consider topological feature representations thereof to evaluate structural connectivity. We perform an ablation study, combined with model interpretability analysis, to evaluate the relevance of the specific image region used for classification. Notably, we find that by choosing a meaningful data representation comprising the left hippocampus, we achieve competitive performance (accuracy 84 ± 7%) comparable to far more complex, heavily parameterised machine learning architectures. This implies that clinical domain knowledge may overrule the importance of model architecture design in the case of AD classification. This opens up new possibilities for interpretable architectures and simplifies model training in terms of computational cost and hardware requirements.
APA
Brüningk, S.C., Hensel, F., Lukas, L.P., Kuijs, M., Jutzeler, C.R. & Rieck, B.. (2021). Back to the basics with inclusion of clinical domain knowledge - A simple, scalable and effective model of Alzheimer’s Disease classification. Proceedings of the 6th Machine Learning for Healthcare Conference, in Proceedings of Machine Learning Research 149:730-754 Available from https://proceedings.mlr.press/v149/bruningk21a.html.

Related Material