Structured and Sparse Canonical Correlation Analysis as a Brain-Wide Multi-Modal Data Fusion Approach
The DOI matches its Crossref record; 42 of the 43 references it cites were found in a registry or an index.
Abstract
Multi-modal data fusion has recently emerged as a comprehensive neuroimaging analysis approach, which usually uses canonical correlation analysis (CCA). However, the current CCA-based fusion approaches face problems like high-dimensionality, multi-collinearity, unimodal feature selection, asymmetry, and loss of spatial information in reshaping the imaging data into vectors. This paper proposes a structured and sparse CCA (ssCCA) technique as a novel CCA method to overcome the above problems. To investigate the performance of the proposed algorithm, we have compared three data fusion techniques: standard CCA, regularized CCA, and ssCCA, and evaluated their ability to detect multi-modal data associations. We have used simulations to compare the performance of these approaches and probe the effects of non-negativity constraint, the dimensionality of features, sample size, and noise power. The results demonstrate that ssCCA outperforms the existing standard and regularized CCA-based fusion approaches. We have also applied the methods to real functional magnetic resonance imaging (fMRI) and structural MRI data of Alzheimer's disease (AD) patients (n = 34) and healthy control (HC) subjects (n = 42) from the ADNI database. The results illustrate that the proposed unsupervised technique differentiates the transition pattern between the subject-course of AD patients and HC subjects with a p-value of less than 1 × 10-6. Furthermore, we have depicted the brain mapping of functional areas that are most correlated with the anatomical changes in AD patients relative to HC subjects.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Regularization and Variable Selection Via the Elastic Net
Zou, H., & Hastie, T. (2005) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.1467-9868.2005.00503.x · Cited by 20,957
RELATIONS BETWEEN TWO SETS OF VARIATES
HOTELLING, H. (1936) · Biometrika · doi:10.1093/biomet/28.3-4.321 · Cited by 5,263
Locality Preserving Projections
He, X., & Niyogi, P. (2003) · Cited by 4,119
Witten, D. M., Tibshirani, R., & Hastie, T. (2009) · Biostatistics · doi:10.1093/biostatistics/kxp008 · Cited by 1,535 · cited 5×
“Therefore, sparse penalties such as l1 -norm are imposed on w 1 and w 2 in the sCCA analysis [10]: where c 1 > 0 and c 2 > 0 control the level of sparsity.”
BackgroundMethodInfluentialAltered functional connectivity in early Alzheimer's disease: A resting‐state fMRI study
Wang, K., Liang, M., Wang, L., Tian, L., Zhang, X., Li, K., & Jiang, T. (2007) · Human Brain Mapping · doi:10.1002/hbm.20324 · Cited by 732 · cited 1×
“The AD group has also shown decreased FC between the left thalamus and inferior frontal gyrus [37] and between prefrontal lobe (such as frontal operculum cortex) and parietal lobes but increased positive correlations within the occipital lobe [38].”
BackgroundSanz-Arigita, E. J., Schoonheim, M. M., Damoiseaux, J. S., Rombouts, S. A. R. B., Maris, E., Barkhof, F., Scheltens, P., & Stam, C. J. (2010) · PLoS ONE · doi:10.1371/journal.pone.0013788 · Cited by 608 · cited 1×
“For example, in [35], it is shown that the AD group has more synchronization between sub-cortical regions such as pallidum and putamen and frontal cortices compared to the HC group.”
ResultPrediction of MCI to AD conversion, via MRI, CSF biomarkers, and pattern classification
Davatzikos, C., Bhatt, P., Shaw, L. M., Batmanghelich, K. N., & Trojanowski, J. Q. (2011) · Neurobiology of Aging · doi:10.1016/j.neurobiolaging.2010.05.023 · Cited by 578 · cited 1×
“caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundEigenvector Centrality Mapping for Analyzing Connectivity Patterns in fMRI Data of the Human Brain
Lohmann, G., Margulies, D. S., Horstmann, A., Pleger, B., Lepsien, J., Goldhahn, D., Schloegl, H., Stumvoll, M., Villringer, A., & Turner, R. (2010) · PLoS ONE · doi:10.1371/journal.pone.0010232 · Cited by 505 · cited 1×
“ECM requires the computation of a voxel-wise connectivity matrix to calculate its eigenvectors [24].”
MethodClass Prediction by Nearest Shrunken Centroids, with Applications to DNA Microarrays
Tibshirani, R., Hastie, T., Narasimhan, B., & Chu, G. (2003) · Statistical Science · doi:10.1214/ss/1056397488 · Cited by 410 · cited 1×
“Treating the covariance matrix as an identity matrix helps us overcome this problem [9], [10], [19].”
BackgroundIdentification of MCI individuals using structural and functional connectivity networks
Wee, C.-Y., Yap, P.-T., Zhang, D., Denny, K., Browndyke, J. N., Potter, G. G., Welsh-Bohmer, K. A., Wang, L., & Shen, D. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2011.10.015 · Cited by 379 · cited 2×
“In data integration approaches, data from different modalities are usually analyzed through separate pipelines and the results combined at the interpretation level to yield decision level fusion [1], [2].”
BackgroundRegularized Generalized Canonical Correlation Analysis
Tenenhaus, A., & Tenenhaus, M. (2011) · Psychometrika · doi:10.1007/s11336-011-9206-8 · Cited by 368
Apostolova, L. G., Green, A. E., Babakchanian, S., Hwang, K. S., Chou, Y.-Y., Toga, A. W., & Thompson, P. M. (2012) · Alzheimer Disease & Associated Disorders · doi:10.1097/wad.0b013e3182163b62 · Cited by 363 · cited 1×
“They include caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundSparse Canonical Correlation Analysis with Application to Genomic Data Integration
Parkhomenko, E., Tritchler, D., & Beyene, J. (2009) · Statistical Applications in Genetics and Molecular Biology · doi:10.2202/1544-6115.1406 · Cited by 343 · cited 4×
“To correlate the simulated datasets, a latent model similar to [9] was applied.”
BackgroundMethodInfluentialLinked independent component analysis for multimodal data fusion
Groves, A. R., Beckmann, C. F., Smith, S. M., & Woolrich, M. W. (2011) · NeuroImage · doi:10.1016/j.neuroimage.2010.09.073 · Cited by 319 · cited 1×
“However, there should be benefits to fuse the data in earlier steps, in particular, after pre-processing but before statistical analysis to yield feature-level fusion [3].”
BackgroundWestman, E., Muehlboeck, J.-S., & Simmons, A. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2012.04.056 · Cited by 302 · cited 2×
“In data integration approaches, data from different modalities are usually analyzed through separate pipelines and the results combined at the interpretation level to yield decision level fusion [1], [2].”
BackgroundDai, Z., Yan, C., Wang, Z., Wang, J., Xia, M., Li, K., & He, Y. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2011.10.003 · Cited by 290 · cited 1×
“Furthermore, compared with the HCs, the AD patients have shown a significant amplitude of lowfrequency fluctuations (ALFF) increases in the hippocampus [39].”
ResultCanonical Correlation Analysis for Data Fusion and Group Inferences
Correa, N., Adali, T., Li, Y.-O., & Calhoun, V. (2010) · IEEE Signal Processing Magazine · doi:10.1109/msp.2010.936725 · Cited by 264 · cited 2×
“Due to this high dimensionality and high noise level, variable selection is critical to avoid over-fitting of the data [6].”
BackgroundInterpretable whole-brain prediction analysis with GraphNet
Grosenick, L., Klingenberg, B., Katovich, K., Knutson, B., & Taylor, J. E. (2013) · NeuroImage · doi:10.1016/j.neuroimage.2012.12.062 · Cited by 201 · cited 1×
“This can make it difficult to interpret features (model will be too sparse) and sensitive to data resampling (e.g., during cross-validation) [42].”
BackgroundWaaijenborg, S., Verselewel de Witt Hamer, P. C., & Zwinderman, A. H. (2008) · Statistical Applications in Genetics and Molecular Biology · doi:10.2202/1544-6115.1329 · Cited by 177
Chen, J., Bushman, F. D., Lewis, J. D., Wu, G. D., & Li, H. (2012) · Biostatistics · doi:10.1093/biostatistics/kxs038 · Cited by 174 · cited 3×
“To facilitate computation, we write constraints on w1 and w2 in Lagrangian form as [10], [16]:”
BackgroundMethodInfluentialCorrea, N. M., Li, Y., Adalı, T., & Calhoun, V. D. (2008) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2008.2008265 · Cited by 153 · cited 1×
“Similar to [5], a CCA-based fusion approach may be used at the feature level.”
MethodSui, J., He, H., Pearlson, G. D., Adali, T., Kiehl, K. A., Yu, Q., Clark, V. P., Castro, E., White, T., Mueller, B. A., et al. (2013) · NeuroImage · doi:10.1016/j.neuroimage.2012.10.051 · Cited by 143 · cited 1×
“modalities but additional benefits may be obtained by combining more than two modalities in a single model and examining N-way data fusion [7].”
BackgroundCorrespondence between fMRI and SNP data by group sparse canonical correlation analysis
Lin, D., Calhoun, V. D., & Wang, Y.-P. (2014) · Medical Image Analysis · doi:10.1016/j.media.2013.10.010 · Cited by 139 · cited 3×
“The groupness constraint in [17] is a special case of our graph-based smoothing constraint.”
BackgroundMethodInfluentialZhou, B., Liu, Y., Zhang, Z., An, N., Yao, H., Wang, P., Wang, L., Zhang, X., & Jiang, T. (2013) · Current Alzheimer Research · doi:10.2174/15672050113109990146 · Cited by 123 · cited 2×
“In addition, increased FC was observed between the bilateral thalamus and inferior temporal gyrus [37].”
BackgroundDu, L., Huang, H., Yan, J., Kim, S., Risacher, S. L., Inlow, M. H., Moore, J. H., Saykin, A. J., Shen, L., & Initiative, A. D. N. (2016) · Bioinformatics · doi:10.1093/bioinformatics/btw033 · Cited by 116 · cited 1×
“A limited number of previous CCA-based fusion approaches investigated this spatial smoothness [14]–[18].”
BackgroundFloch, É. L., Guillemot, V., Frouin, V., Pinel, P., Lalanne, C., Trinchera, L., Tenenhaus, A., Moreno, A., Zilbovicius, M., Bourgeron, T., et al. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2012.06.061 · Cited by 114 · cited 1×
“performed for each modality separately [8], which is blind”
Wink, A. M., de Munck, J. C., van der Werf, Y. D., van den Heuvel, O. A., & Barkhof, F. (2012) · Brain Connectivity · doi:10.1089/brain.2012.0087 · Cited by 111 · cited 1×
“Eigenvector centrality mapping (ECM) of the rs-fMRI time series in the MNI-space was performed using the fast ECM (fECM) software [23], which yields a voxel-wise measure of the centrality of a node (voxel) to the functional brain network.”
MethodLi, J., Pan, P., Huang, R., & Shang, H. (2011) · Neuroscience & Biobehavioral Reviews · doi:10.1016/j.neubiorev.2011.12.001 · Cited by 107 · cited 1×
“They include caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundYao, H., Liu, Y., Zhou, B., Zhang, Z., An, N., Wang, P., Wang, L., Zhang, X., & Jiang, T. (2013) · European Journal of Radiology · doi:10.1016/j.ejrad.2013.03.019 · Cited by 97 · cited 1×
“Also, in [36], it is shown that the FC between the amygdala and the default mode network (DMN) in the AD group is impaired.”
BackgroundHarasty, J. A., Halliday, G. M., Kril, J. J., & Code, C. (1999) · Brain · doi:10.1093/brain/122.4.675 · Cited by 96 · cited 1×
“caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundStereological quantification of the cerebellum in patients with Alzheimer's disease
Andersen, K., Andersen, B. B., & Pakkenberg, B. (2010) · Neurobiology of Aging · doi:10.1016/j.neurobiolaging.2010.06.013 · Cited by 84 · cited 1×
“caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundStructured Sparse Canonical Correlation Analysis
Chen, X., Liu, H., & Carbonell, J. G. (2018) · doi:10.1184/r1/6473711 · Cited by 81
Avants, B. B., Libon, D. J., Rascovsky, K., Boller, A., McMillan, C. T., Massimo, L., Coslett, H. B., Chatterjee, A., Gross, R. G., & Grossman, M. (2014) · NeuroImage · doi:10.1016/j.neuroimage.2013.09.048 · Cited by 78 · cited 2×
“Because existence of negative weights in the canonical coefficients makes the interpretations difficult, it seems rational to add a simple non-negativity restriction as a regularization term to the standard CCA [11], [12].”
BackgroundMethodSegmentation and volumetric analysis of the caudate nucleus in Alzheimer's disease
Jiji, S. G., Smitha, K. A., Gupta, A. K., Pillai, V. P. M., & Jayasree, R. S. (2013) · European Journal of Radiology · doi:10.1016/j.ejrad.2013.03.012 · Cited by 71 · cited 1×
“They include caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundStriatal morphology in early-onset and late-onset Alzheimer's disease: a preliminary study
Pievani, M., Bocchetta, M., Boccardi, M., Cavedo, E., Bonetti, M., Thompson, P. M., & Frisoni, G. B. (2013) · Neurobiology of Aging · doi:10.1016/j.neurobiolaging.2013.01.016 · Cited by 62 · cited 1×
“They include caudate [27], hippocampus [1], lateral ventricle [28], thalamus [2], planum polare [29], planum temporale [29], heschl’s gyrus [30], cerebral WM [31], accumbens [32], amygdala [1], [2], [29], paracingulate gyrus [33], and cerebellum [34].”
BackgroundDeleus, F., & Hulle, M. M. V. (2011) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2010.11.029 · Cited by 44 · cited 2×
“Therefore, canonical coefficients associated with these voxels should have similar magnitudes to reflect the underlying association [12], [13].”
BackgroundSparse and Functional Principal Components Analysis
Allen, G. I., & Weylandt, M. (2019) · IEEE Data Science Workshop (DSW) · doi:10.1109/dsw.2019.8755778 · Cited by 36
Voxel-level functional connectivity using spatial regularization
Baldassano, C. A., Iordan, M. C., Beck, D. M., & Fei-Fei, L. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2012.07.046 · Cited by 34 · cited 1×
“Therefore, canonical coefficients associated with these voxels should have similar magnitudes to reflect the underlying association [12], [13].”
BackgroundAltered Functional Connectivity of the Marginal Division in Alzheimer’s Disease
Zhang, Z., Liu, Y., Zhou, B., Zheng, J., Yao, H., An, N., Wang, P., Guo, Y., Dai, H., Wang, L., et al. (2014) · Current Alzheimer Research · doi:10.2174/1567205011666140110112608 · Cited by 22 · cited 1×
“Moreover, compared with the HCs, a decreased FC pattern was observed between the marginal division and the amygdala/parahippocampal region, the inferior frontal gyrus and the cerebellum for the AD patients [40].”
Structured sparse CCA for brain imaging genetics via graph OSCAR
Initiative, A. D. N., Du, L., Huang, H., Yan, J., Kim, S., Risacher, S. L., Inlow, M. H., Moore, J. H., Saykin, A. J., & Shen, L. (2016) · BMC Systems Biology · doi:10.1186/s12918-016-0312-1 · Cited by 14
Mohammadi‐Nejad, A., Hossein‐Zadeh, G., & Hamid Soltanian-zadeh, A. (2016) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi.2016.7493392 · Cited by 4 · cited 1×
“Also in [26], we probed the performance of the fusion methods using a low dimensional simulated dataset.”
MethodSelf-citationDoré, V., Bourgeat, P., Fripp, J., Acosta, O., Chételat, G., Szoeke, C. E. I., Ellis, K. A., Martins, R. N., Villemagne, V. L., Masters, C. L., et al. (2012) · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · doi:10.1117/12.911752 · Cited by 2
Cited by
Zhou, T., Liu, M., Thung, K.-H., & Shen, D. (2019) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2019.2913158 · Cited by 226
“In particular, the fusion of multi-modality data has advanced the progress of automatic AD diagnosis, thanks to the complementary information contained within them [12]–[16].”
A technical review of canonical correlation analysis for neuroscience applications
Zhuang, X., Yang, Z., & Cordes, D. (2020) · Human Brain Mapping · doi:10.1002/hbm.25090 · Cited by 209
“and to link structural and functional brain activities (Lisowska & Rekik, 2019; Mohammadi-Nejad et al., 2017).”
BackgroundA Survey on Canonical Correlation Analysis
Yang, X., Liu, W., Liu, W., & Tao, D. (2021) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2019.2958342 · Cited by 202
“Mohammadi et al. [100] investigated the the multi-modal data fusion problem by the structured and sparse CCA (ssCCA), which adds smoothing constraints to preserve image spatial structure.”
BackgroundMulti-Modal Graph Learning for Disease Prediction
Zheng, S., Zhu, Z., Liu, Z., Guo, Z., Liu, Y., Yang, Y., & Zhao, Y. (2022) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2022.3159264 · Cited by 196
Liu, L., Chen, S., Zhang, F., Wu, F., Pan, Y. F., & Wang, J. (2019) · Neural Computing and Applications · doi:10.1007/s00521-019-04096-x · Cited by 112
Metadata
- Authors
- Mohammadi-Nejad, A.-R., Hossein-Zadeh, G.-A., & Soltanian-Zadeh, H.
- Venue
- IEEE Transactions on Medical Imaging
- Published
- 2017-07
- Volume
- 36
- Issue
- 7
- Pages
- 1438-1448
- ISSN
- 0278-0062 (ISSN-L)
- 1558-254X
- OpenAlex
- W2597268455
- Cited by
- 56
- References
- 43
- Citations by year
- 2026: 4 · 2025: 4 · 2024: 6 · 2023: 6 · 2022: 11 · 2021: 5 · 2020: 8 · 2019: 6 · 2018: 3 · 2017: 3
- FWCI
- 3.10
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE Transactions on Medical Imaging (publishedVersion)
- PubMed (publishedVersion)
- Funders
- Iranian Cognitive Science and Technologies Council Research (108)
- Alzheimer’s Disease Neuroimaging Initiative (ADNI)
- National Institute on Aging and the National Institute of Biomedical Imaging and Bioengineering, National Institutes of Health (U01AG024904)
- Department of Defense (W81XWH-12-2-0012)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Functional Brain Connectivity Studies
- Physical Sciences › Mathematics › Computational Mathematics › Tensor decomposition and applications
- Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
- Keywords
- multimodal fusion, functional MRI, structural MRI, Alzheimer's disease, nonnegativity constraints, high dimensionality, multicollinearity, unsupervised feature selection, brain mapping, ADNI
- MeSH terms
- Algorithms, Brain, Brain Mapping, Humans, Magnetic Resonance Imaging, Neuroimaging
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.