Multimodal Subspace Independent Vector Analysis (MSIVA)

This repository contains MATLAB implementation for Multimodal Subspace Independent Vector Analysis (MSIVA).

Workflows

Description	Script
MSIVA default initialization	run_mgpca_ica.m
Unimodal initialization	run_pca_ica.m
Multimodal initialization	run_mgpca_gica.m

Experiments

Description	Script
Synthetic data experiment	func_sim.m
Neuroimaging data experiment	func_img.m

Figures

Imaging Neuroscience

Figure	Script
Fig. 1	plot_subspace_struct.ipynb
Fig. 3	plot_sim.ipynb
Fig. 4	plot_sim.ipynb
Fig. 5a	plot_img_ukb.ipynb
Fig. 5b	plot_img_sz.ipynb
Fig. 6	plot_img_loss.ipynb
Fig. 7	plot_img_ukb.ipynb
Fig. 8	plot_img_sz.ipynb
Fig. 9	dualcodeImage_AY_geomedian.m, plot_img_ukb.ipynb
Fig. 10	dualcodeImage_AY_geomedian.m, plot_img_sz.ipynb
Fig. 11	plot_sig_voxel.ipynb
Fig. 12	(1) Run age_delta.m to compute brain-age delta. (2) Run compute_geometric_median.py to compute geometric median of brain-age delta. (3) Run phenotype_map.py to compute spatial correlation between brain-age delta and phenotype variable. (4) Use dualcodeImage_beta1.m, dualcodeImage_delta2p_std.m, dualcodeImage_delta2p_geomedian.m, and dualcodeImage_phenotype.m to plot the dual-coded maps.
Fig. S1	dualcodeImage_beta1.m
Fig. S2	plot_init_corr.ipynb
Figs. S3 & S4	plot_loss.ipynb
Fig. S5	plot_itc.ipynb
Fig. S6a	plot_img_ukb_rdc.ipynb
Fig. S6b	plot_img_sz_rdc.ipynb
Fig. S7	plot_img_sz.ipynb
Fig. S8	dualcodeImage_AY_geomedian.m, plot_img_ukb.ipynb
Fig. S9	dualcodeImage_AY_geomedian.m, plot_img_sz.ipynb
Fig. S10	plot_num_crossmodal_voxel.ipynb
Figs. S11 & S12	compare_mmiva_msiva_ukb.ipynb
Figs. S13 & S14	compare_mmiva_msiva_sz.ipynb

ISBI

Figure	Script
Fig. 1	plot_subspace_struct.ipynb
Fig. 2	plot_sim.ipynb
Fig. 3	plot_sim.ipynb
Fig. 4	plot_img.ipynb
Fig. 5	plot_img.ipynb
Fig. 6	dualmap.m

Prerequisites

MSIVA builds on Multidataset Independent Subspace Analysis (MISA), which is already included in this repository. It also requires:

• MATLAB Optimization Toolbox
• MATLAB Statistics and Machine Learning Toolbox
• Group ICA of fMRI Toolbox (GIFT) — see installation instructions below.

Installing GIFT

1. Clone the repository (in a terminal):

git clone https://github.com/trendscenter/gift.git

2. Navigate to the GIFT directory (in MATLAB):

cd gift/GroupICAT/icatb

3. Run the installer (in the MATLAB command window):

groupica fmri 

A GUI window will open upon successful installation — you can close it by clicking Exit.

References

If you find this repository useful, please cite the following papers:

@article{li2026multimodal,
    author = {Li, Xinhui and Kochunov, Peter and Adali, Tulay and Silva, Rogers F. and Calhoun, Vince D.},
    title = {Multimodal subspace independent vector analysis effectively captures latent relationships between brain structure and function},
    journal = {Imaging Neuroscience},
    year = {2026},
    month = {05},
    issn = {2837-6056},
    doi = {10.1162/IMAG.a.1266},
    url = {https://doi.org/10.1162/IMAG.a.1266},
    eprint = {https://direct.mit.edu/imag/article-pdf/doi/10.1162/IMAG.a.1266/2600396/imag.a.1266.pdf},
}

@INPROCEEDINGS{li2023multimodal,
  author={Li, Xinhui and Adali, Tulay and Silva, Rogers F. and Calhoun, Vince D.},
  booktitle={2023 IEEE 20th International Symposium on Biomedical Imaging (ISBI)}, 
  title={Multimodal Subspace Independent Vector Analysis Better Captures Hidden Relationships in Multimodal Neuroimaging Data}, 
  year={2023},
  pages={1-5},
  doi={10.1109/ISBI53787.2023.10230605}
}