Skip to main content

Advertisement

Log in

Deep Learning‐based Classification of Resting‐state fMRI Independent‐component Analysis

  • Original Article
  • Published:
Neuroinformatics Aims and scope Submit manuscript

Abstract

Functional connectivity analyses of fMRI data have shown that the activity of the brain at rest is spatially organized into resting-state networks (RSNs). RSNs appear as groups of anatomically distant but functionally tightly connected brain regions. Inter-RSN intrinsic connectivity analyses may provide an optimal spatial level of integration to analyze the variability of the functional connectome. Here we propose a deep learning approach to enable the automated classification of individual independent-component (IC) decompositions into a set of predefined RSNs. Two databases were used in this work, BIL&GIN and MRi-Share, with 427 and 1811 participants, respectively. We trained a multilayer perceptron (MLP) to classify each IC as one of 45 RSNs, using the IC classification of 282 participants in BIL&GIN for training and a 5-dimensional parameter grid search for hyperparameter optimization. It reached an accuracy of 92 %. Predictions for the remaining individuals in BIL&GIN were tested against the original classification and demonstrated good spatial overlap between the cortical RSNs. As a first application, we created an RSN atlas based on MRi-Share. This atlas defined a brain parcellation in 29 RSNs covering 96 % of the gray matter. Second, we proposed an individual-based analysis of the subdivision of the default-mode network into 4 networks. Minimal overlap between RSNs was found except in the angular gyrus and potentially in the precuneus. We thus provide the community with an individual IC classifier that can be used to analyze one dataset or to statistically compare different datasets for RSN spatial definitions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8

Similar content being viewed by others

Data Availability

Both BIL&GIN and MRi-Share data are open to collaborations and partnerships and support local, national and international collaborations from the public or private sector. For BIL&GIN, requests to access the database should be sent through the GIN website https://www.gin.cnrs.fr/en/current-research/axis2/bilgin-en/. For MRi-Share, requests to access the database should be sent to contact@i-share.fr.

References

Download references

Acknowledgements

This work has been funded by a governmental grant of “Agence Nationale de la Recherche (ANR)” (Ginesilab, ANR-16-LCV2-0006), a University Bordeaux/CEA/CNRS and Fealinx joint laboratory. The i-Share cohort has been funded by a grant ANR-10-COHO-05-01 as part of the “Programme pour les Investissements d’Avenir”). Supplementary funding was received from the “Conseil Régional de Nouvelle-Aquitaine”, reference 4370420. The MRi-Share cohort has been supported by ANR-10-LABX-57 (TRAIL). Computer time for the MRi-Share study was provided by the MCIA (Mésocentre de Calcul Intensif Aquitain) of the Université de Bordeaux and the Université de Pau et des Pays de l’Adour.

Funding

This work has been funded by a grant from “Agence Nationale de la Recherche” (Ginesilab, ANR-16-LCV2-0006), a University Bordeaux/CEA/CNRS and Fealinx joint laboratory. The i-Share cohort has been funded by a grant ANR-10-COHO-05-01 as part of the “Programme pour les Investissements d’Avenir”). Supplementary funding was received from the “Conseil Régional of Nouvelle-Aquitaine”, reference 4370420. The MRi-Share cohort has been supported by ANR-10-LABX-57 (TRAIL).

Author information

Authors and Affiliations

Authors

Contributions

PB, MJ designed the study. PB, VV, M-FG, PYH, MJ designed the code and ran the analysis. BM (BIL&GIN and MRi-Share) and CT (iShare) supervised the acquisition of the cohorts. All authors participated in the manuscript writing.

Corresponding author

Correspondence to Marc Joliot.

Ethics declarations

Conflicts of Interest/Competing Interests

Not applicable.

Ethics Approval

The BIL&GIN study was approved by the ethics committee of Basse-Normandie (France). The MRi-Share study was approved by the ethics committee of Bordeaux (France). 

Consent to Participate

Participants in the BIL&GIN and MRi-Share provided their written consent for participation in the study.

Consent for Publication

All authors have agreed to the publication of the results presented in this work.

Code Availability

The classifier and documentation are provided for free upon request to the author of the publication.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

ESM 1

(DOCX 322 KB)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Nozais, V., Boutinaud, P., Verrecchia, V. et al. Deep Learning‐based Classification of Resting‐state fMRI Independent‐component Analysis. Neuroinform 19, 619–637 (2021). https://doi.org/10.1007/s12021-021-09514-x

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12021-021-09514-x

Keywords

Navigation