Abstract
Learning biological markers for a specific brain pathology is often impaired by the size of the dataset. With the advent of large open datasets in the general population, new learning strategies have emerged. In particular, deep representation learning consists of training a model via pretext tasks that can be used to solve downstream clinical problems of interest. More recently, self-supervised learning provides a rich framework for learning representations by contrasting transformed samples. These methods rely on carefully designed data manipulation to create semantically similar but syntactically different samples. In parallel, domain-specific architectures such as spherical convolutional neural networks can learn from cortical brain measures in order to reveal original biomarkers. Unfortunately, only a few surface-based augmentations exist, and none of them have been applied in a self-supervised learning setting. We perform experiments on two open source datasets: Big Healthy Brain and Healthy Brain Network. We propose new augmentations for the cortical brain: baseline augmentations adapted from classical ones for training convolutional neural networks, typically on natural images, and new augmentations called MixUp. The results suggest that surface-based self-supervised learning performs comparably to supervised baselines, but generalizes better to different tasks and datasets. In addition, the learned representations are improved by the proposed MixUp augmentations. The code is available on GitHub (https://github.com/neurospin-projects/2022_cambroise_surfaugment).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Alexander, L.M., et al.: An open resource for transdiagnostic research in pediatric mental health and learning disorders. Sci. Data 4, 170181 (2017)
Alzubaidi, L., et al.: Towards a better understanding of transfer learning for medical imaging: a case study. Appl. Sci. 10(13), 4523 (2020)
Bommasani, R., et al.: On the opportunities and risks of foundation models (2021)
Caron, M., et al.: Emerging properties in self-supervised vision transformers. In: ICCV (2021)
Chen, T., Kornblith, S., Norouzi, M., Hinton, G.: A simple framework for contrastive learning of visual representations. In: ICML, pp. 1597–1607 (2020)
Cohen, T.S., Geiger, M., Köhler, J., Welling, M.: Spherical CNNs. In: ICLR (2018)
Dadi, K., Varoquaux, G., Houenou, J., Bzdok, D., Thirion, B., Engemann, D.: Population modeling with machine learning can enhance measures of mental health. GigaScience 10(10), giab071 (2021)
Defferrard, M., Bresson, X., Vandergheynst, P.: Convolutional neural networks on graphs with fast localized spectral filtering. In: NeurIPS, vol. 29 (2016)
Defferrard, M., Milani, M., Gusset, F., Perraudin, N.: DeepSphere: a graph-based spherical CNN. In: ICLR (2020)
Dufumier, B., et al.: Contrastive learning with continuous proxy meta-data for 3D MRI classification. In: de Bruijne, M., et al. (eds.) MICCAI 2021. LNCS, vol. 12902, pp. 58–68. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87196-3_6
Dufumier, B., Grigis, A., Victor, J., Ambroise, C., Frouin, V., Duchesnay, E.: OpenBHB: a large-scale multi-site brain MRI data-set for age prediction and debiasing. Neuroimage 263, 119637 (2022)
Eitel, F., Schulz, M.A., Seiler, M., Walter, H., Ritter, K.: Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research. Exp. Neurol. 339, 113608 (2021)
Fischl, B., Sereno, M.I., Dale, A.M.: Cortical surface-based analysis: II: inflation, flattening, and a surface-based coordinate system. Neuroimage 9(2), 195–207 (1999)
Fortin, J.P., et al.: Harmonization of multi-site diffusion tensor imaging data. Neuroimage 161, 149–170 (2017)
Greve, D.N., et al.: A surface-based analysis of language lateralization and cortical asymmetry. J. Cogn. Neurosci. 25(9), 1477–1492 (2013)
Grill, J.B., et al.: Bootstrap your own latent - a new approach to self-supervised learning. In: NeurIPS, vol. 33 (2020)
He, K., Fan, H., Wu, Y., Xie, S., Girshick, R.: Momentum contrast for unsupervised visual representation learning. In: CVPR (2020)
Jiang, C.M., Huang, J., Kashinath, K., Prabhat, Marcus, P., Niessner, M.: Spherical CNNs on unstructured grids. In: ICLR (2019)
Khosla, P., et al.: Supervised contrastive learning. In: NeurIPS (2020)
Kipf, T.N., Welling, M.: Semi-supervised classification with graph convolutional networks. In: ICLR (2017)
Raghu, M., Zhang, C., Kleinberg, J., Bengio, S.: Transfusion: understanding transfer learning for medical imaging. In: NeurIPS, vol. 32 (2019)
Saunshi, N., et al.: Understanding contrastive learning requires incorporating inductive biases. In: ICML, vol. 162, pp. 19250–19286 (2022)
Seong, S.B., Pae, C., Park, H.J.: Geometric convolutional neural network for analyzing surface-based neuroimaging data. Front. Neuroinform. 12, 42 (2018)
Tian, Y., Sun, C., Poole, B., Krishnan, D., Schmid, C., Isola, P.: What makes for good views for contrastive learning? In: Advances in Neural Information Processing Systems, vol. 33, pp. 6827–6839. Curran Associates, Inc. (2020)
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Lió, P., Bengio, Y.: Graph attention networks. In: ICLR (2018)
Wang, F., Liu, H.: Understanding the behaviour of contrastive loss. In: IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2021, virtual, 19–25 June 2021, pp. 2495–2504 (2021)
Yoon, J., Zhang, Y., Jordon, J., van der Schaar, M.: VIME: extending the success of self- and semi-supervised learning to tabular domain. In: NeurIPS, vol. 33 (2020)
Zbontar, J., Jing, L., Misra, I., LeCun, Y., Deny, S.: Barlow Twins: self-supervised learning via redundancy reduction. In: ICML, vol. 139 (2021)
Zhai, X., Oliver, A., Kolesnikov, A., Beyer, L.: S4L: self-supervised semi-supervised learning. In: ICCV (2019)
Zhang, H., Cissé, M., Dauphin, Y.N., Lopez-Paz, D.: mixup: beyond empirical risk minimization. In: 6th International Conference on Learning Representations, ICLR 2018, Conference Track Proceedings, Vancouver, BC, Canada, 30 April–3 May 2018 (2018)
Zhao, F., et al.: Spherical deformable U-Net: application to cortical surface parcellation and development prediction. IEEE Trans. Med. Imaging 40, 1217–1228 (2021)
Zhao, F., et al.: Spherical U-Net on cortical surfaces: methods and applications. In: IPMI (2019)
Data Use Declaration and Acknowledgment
The datasets analyzed during the current study are available online: OpenBHB in IEEEDataPort (doi 10.21227/7jsg-jx57), and HBN in NITRC (Release 10).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
1 Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Ambroise, C., Frouin, V., Dufumier, B., Duchesnay, E., Grigis, A. (2023). MixUp Brain-Cortical Augmentations in Self-supervised Learning. In: Abdulkadir, A., et al. Machine Learning in Clinical Neuroimaging. MLCN 2023. Lecture Notes in Computer Science, vol 14312. Springer, Cham. https://doi.org/10.1007/978-3-031-44858-4_10
Download citation
DOI: https://doi.org/10.1007/978-3-031-44858-4_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-44857-7
Online ISBN: 978-3-031-44858-4
eBook Packages: Computer ScienceComputer Science (R0)