Transport on Riemannian manifold for functional connectivity-based classification
- PMID: 25485405
- DOI: 10.1007/978-3-319-10470-6_51
Transport on Riemannian manifold for functional connectivity-based classification
Abstract
We present a Riemannian approach for classifying fMRI connectivity patterns before and after intervention in longitudinal studies. A fundamental difficulty with using connectivity as features is that covariance matrices live on the positive semi-definite cone, which renders their elements inter-related. The implicit independent feature assumption in most classifier learning algorithms is thus violated. In this paper, we propose a matrix whitening transport for projecting the covariance estimates onto a common tangent space to reduce the statistical dependencies between their elements. We show on real data that our approach provides significantly higher classification accuracy than directly using Pearson's correlation. We further propose a non-parametric scheme for identifying significantly discriminative connections from classifier weights. Using this scheme, a number of neuroanatomically meaningful connections are found, whereas no significant connections are detected with pure permutation testing.
Similar articles
-
Transport on Riemannian Manifold for Connectivity-Based Brain Decoding.IEEE Trans Med Imaging. 2016 Jan;35(1):208-16. doi: 10.1109/TMI.2015.2463723. Epub 2015 Aug 3. IEEE Trans Med Imaging. 2016. PMID: 26259016
-
Discriminative sparse connectivity patterns for classification of fMRI Data.Med Image Comput Comput Assist Interv. 2014;17(Pt 3):193-200. doi: 10.1007/978-3-319-10443-0_25. Med Image Comput Comput Assist Interv. 2014. PMID: 25320799 Free PMC article.
-
Structurally-informed Bayesian functional connectivity analysis.Neuroimage. 2014 Feb 1;86:294-305. doi: 10.1016/j.neuroimage.2013.09.075. Epub 2013 Oct 10. Neuroimage. 2014. PMID: 24121202
-
Automatic sleep staging using fMRI functional connectivity data.Neuroimage. 2012 Oct 15;63(1):63-72. doi: 10.1016/j.neuroimage.2012.06.036. Epub 2012 Jun 26. Neuroimage. 2012. PMID: 22743197
-
Bayesian networks for fMRI: a primer.Neuroimage. 2014 Feb 1;86:573-82. doi: 10.1016/j.neuroimage.2013.10.020. Epub 2013 Oct 18. Neuroimage. 2014. PMID: 24140939 Review.
Cited by
-
Tangent functional connectomes uncover more unique phenotypic traits.iScience. 2023 Aug 12;26(9):107624. doi: 10.1016/j.isci.2023.107624. eCollection 2023 Sep 15. iScience. 2023. PMID: 37694156 Free PMC article.
-
Multiscale functional connectivity patterns of the aging brain learned from harmonized rsfMRI data of the multi-cohort iSTAGING study.Neuroimage. 2023 Apr 1;269:119911. doi: 10.1016/j.neuroimage.2023.119911. Epub 2023 Jan 30. Neuroimage. 2023. PMID: 36731813 Free PMC article.
-
Discovering functional connectivity features characterizing multiple sclerosis phenotypes using explainable artificial intelligence.Hum Brain Mapp. 2023 Apr 15;44(6):2294-2306. doi: 10.1002/hbm.26210. Epub 2023 Jan 30. Hum Brain Mapp. 2023. PMID: 36715247 Free PMC article.
-
Isolation of subjectively reported sleepiness and objectively measured vigilance during sleep deprivation: a resting-state fMRI study.Cogn Neurodyn. 2022 Oct;16(5):1151-1162. doi: 10.1007/s11571-021-09772-0. Epub 2022 Jan 27. Cogn Neurodyn. 2022. PMID: 36237404 Free PMC article.
-
Riemannian Geometry of Functional Connectivity Matrices for Multi-Site Attention-Deficit/Hyperactivity Disorder Data Harmonization.Front Neuroinform. 2022 May 23;16:769274. doi: 10.3389/fninf.2022.769274. eCollection 2022. Front Neuroinform. 2022. PMID: 35685944 Free PMC article.
MeSH terms
LinkOut - more resources
Medical