Advanced search
1 file | 5.37 MB Add to list

Counterfactual functional connectomes for neurological classifier selection

Author
Organization
Project
Abstract
Functional connectivity expresses the correlation of brain activity between regions and helps in understanding and diagnosing neurological conditions and disorders. It also provides discriminative features for machine learning classifiers. We propose a model-agnostic method that produces realistic counterfactual functional connectomes by altering the posterior distribution of a hierarchical variational auto-encoder and de-noising the result. We evaluate our method on three autism spectrum disorder classifiers for resting state fMRI. The generated counterfactuals include plausible changes in line with medical literature and the brain's functional anatomy. Our approach strives for explainability and collaboration with medical experts, starting from the model selection.

Downloads

  • fmri counterfactuals paper-8.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 5.37 MB

Citation

Please use this url to cite or link to this publication:

MLA
Vercheval, Nicolas, et al. “Counterfactual Functional Connectomes for Neurological Classifier Selection.” 2023 31st European Signal Processing Conference (EUSIPCO), IEEE, 2023, pp. 1050–54.
APA
Vercheval, N., Benčević, M., Mužević, D., Galić, I., & Pizurica, A. (2023). Counterfactual functional connectomes for neurological classifier selection. 2023 31st European Signal Processing Conference (EUSIPCO), 1050–1054. IEEE.
Chicago author-date
Vercheval, Nicolas, Marin Benčević, Dario Mužević, Irena Galić, and Aleksandra Pizurica. 2023. “Counterfactual Functional Connectomes for Neurological Classifier Selection.” In 2023 31st European Signal Processing Conference (EUSIPCO), 1050–54. IEEE.
Chicago author-date (all authors)
Vercheval, Nicolas, Marin Benčević, Dario Mužević, Irena Galić, and Aleksandra Pizurica. 2023. “Counterfactual Functional Connectomes for Neurological Classifier Selection.” In 2023 31st European Signal Processing Conference (EUSIPCO), 1050–1054. IEEE.
Vancouver
1.
Vercheval N, Benčević M, Mužević, D, Galić I, Pizurica A. Counterfactual functional connectomes for neurological classifier selection. In: 2023 31st European Signal Processing Conference (EUSIPCO). IEEE; 2023. p. 1050–4.
IEEE
[1]
N. Vercheval, M. Benčević, D. Mužević, I. Galić, and A. Pizurica, “Counterfactual functional connectomes for neurological classifier selection,” in 2023 31st European Signal Processing Conference (EUSIPCO), Helsinki, Finland, 2023, pp. 1050–1054.
@inproceedings{01H961WDD8EVT64BK8ZJTXYPYB,
  abstract     = {{Functional connectivity expresses the correlation of brain activity between regions and helps in understanding and diagnosing neurological conditions and disorders. It also provides discriminative features for machine learning classifiers. We propose a model-agnostic method that produces realistic counterfactual functional connectomes by altering the posterior distribution of a hierarchical variational auto-encoder and de-noising the result. We evaluate our method on three autism spectrum disorder classifiers for resting state fMRI. The generated counterfactuals include plausible changes in line with medical literature and the brain's functional anatomy. Our approach strives for explainability and collaboration with medical experts, starting from the model selection.}},
  author       = {{Vercheval, Nicolas and Benčević, Marin and Mužević,, Dario and Galić, Irena and Pizurica, Aleksandra}},
  booktitle    = {{2023 31st European Signal Processing Conference (EUSIPCO)}},
  isbn         = {{9789464593600}},
  language     = {{eng}},
  location     = {{Helsinki, Finland}},
  pages        = {{1050--1054}},
  publisher    = {{IEEE}},
  title        = {{Counterfactual functional connectomes for neurological classifier selection}},
  year         = {{2023}},
}