Data#
DIPY provides access to several datasets used for testing, tutorials, and research. All datasets are downloaded automatically from the internet and cached locally on your machine for future use.
How to get data#
The list of datasets can be retrieved using:
from dipy.workflows.io import FetchFlow
available_data = FetchFlow.get_fetcher_datanames().keys()
To retrieve all datasets, the following workflow can be run:
from tempfile import TemporaryDirectory
from dipy.workflows.io import FetchFlow
fetch_flow = FetchFlow()
with TemporaryDirectory() as out_dir:
fetch_flow.run(['all'], out_dir=out_dir)
If you want to download a particular dataset, you can do:
from tempfile import TemporaryDirectory
from dipy.workflows.io import FetchFlow
fetch_flow = FetchFlow()
with TemporaryDirectory() as out_dir:
fetch_flow.run(['bundle_fa_hcp'], out_dir=out_dir)
or:
from dipy.data import fetch_bundle_fa_hcp
files, folder = fetch_bundle_fa_hcp()
Datasets List#
Details about datasets available in DIPY are described in the table below:
Name |
Synthetic/Phantom/Human/Animal |
Data features (structural; diffusion; label information) |
Scanner |
DIPY name |
Citations |
|---|---|---|---|---|---|
Tractogram file formats examples |
Synthetic |
Tractogram file formats (.dpy, .fib, .tck, .trk) |
bundle_file_formats_example |
Rheault, F. (2019). Bundles for tractography file format testing and example (Version 1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.3352379 |
|
CENIR HCP-like dataset |
Multi-shell data: b-vals: [200, 400, 1000, 2000, 3000] (s/mm^2); [20, 20, 202, 204, 206] gradient directions; Corrected for Eddy currents |
cenir_multib |
|||
CFIN dataset |
T1; Multi-shell data: b-vals: [200, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, 2000, 2200, 2400, 2600, 2800, 3000] (s/mm^2); 496 gradient directions |
cfin_multib |
Hansen, B., Jespersen, S.. Data for evaluation of fast kurtosis strategies, b-value optimization and exploration of diffusion MRI contrast. Sci Data 3, 160072 (2016). doi:10.1038/sdata.2016.72 |
||
Gold standard streamlines IO testing |
Synthetic |
Tractogram file formats (.dpy, .fib, .tck, .trk) |
gold_standard_io |
Rheault, F. (2019). Gold standard for tractogram io testing (Version 1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.2651349 |
|
HCP842 bundle atlas |
Human |
Whole brain/bundle-wise tractograms in MNI space; 80 bundles |
Human Connectome Project (HCP) scanner |
bundle_atlas_hcp842 |
Garyfallidis, E., et al. Recognition of white matter bundles using local and global streamline-based registration and clustering. NeuroImage 170 (2017): 283-297; Yeh, F.-C., et al. Population-averaged atlas of the macroscale human structural connectome and its network topology. NeuroImage 178 (2018): 57-68. <a href=’https://figshare.com/articles/Advanced_Atlas_of_80_Bundles_in_MNI_space/7375883’>figshare.com/articles/Advanced_Atlas_of_80_Bundles_in_MNI_space/7375883</a> |
HCP bundle FA |
Human |
Fractional Anisotropy (FA); 2 bundles |
bundle_fa_hcp |
||
HCP tractogram |
Human |
Whole brain tractogram |
Human Connectome Project (HCP) scanner |
target_tractogram_hcp |
|
ISBI 2013 |
Phantom |
Multi shell data: b-vals: [0, 1500, 2500] (s/mm^2); 64 gradient directions |
isbi2013_2shell |
Daducci, A., et al. Quantitative Comparison of Reconstruction Methods for Intra-Voxel Fiber Recovery From Diffusion MRI. IEEE Transactions on Medical Imaging, vol. 33, no. 2, pp. 384-399, Feb. 2014. <a href=’https://hardi.epfl.ch/static/events/2013_ISBI/testing_data.html’>HARDI reconstruction challenge 2013</a> |
|
IVIM dataset |
Human |
Multi shell data: b-vals: [0, 10, 20, 30, 40, 60, 80, 100, 120, 140, 160, 180, 200, 300, 400, 500, 600, 700, 800, 900, 1000] (s/mm^2); 21 gradient directions |
fetch_ivim |
Peterson, Eric (2016): IVIM dataset. figshare. Dataset. <a href=’https://doi.org/10.6084/m9.figshare.3395704.v1’>figshare.com/articles/dataset/IVIM_dataset/3395704/1</a> |
|
MNI template |
Human |
MNI 2009a T1, T2; 2009c T1, T1 mask |
mni_template |
Fonov, V.S., Evans, A.C., Botteron, K., Almli, C.R., McKinstry, R.C., Collins, D.L., BDCG. Unbiased average age-appropriate atlases for pediatric studies. NeuroImage, Volume 54, Issue 1, January 2011, ISSN 1053–8119, doi:10.1016/j.neuroimage.2010.07.033; Fonov, V.S., Evans, A.C., McKinstry, R.C., Almli, C.R., Collins, D.L. Unbiased nonlinear average age-appropriate brain templates from birth to adulthood, NeuroImage, Volume 47, Supplement 1, July 2009, Page S102 Organization for Human Brain Mapping 2009 Annual Meeting, doi:10.1016/S1053-8119(09)70884-5 <a href=’https://www.bic.mni.mcgill.ca/ServicesAtlases/ICBM152NLin2009’>ICBM 152 Nonlinear atlases version 2009</a> |
|
qt-dMRI C57Bl6 mice dataset |
Animal |
2 C57Bl6 mice test-retest qt-dMRI; Corpus callosum (CC) bundle masks |
qtdMRI_test_retest_2subjects |
Wassermann, D., Santin, M., Philippe, A.-C., Fick, R., Deriche, R., Lehericy, S., Petiet, A. (2017). Test-Retest qt-dMRI datasets for “Non-Parametric GraphNet-Regularized Representation of dMRI in Space and Time” [Data set]. Zenodo. https://doi.org/10.5281/zenodo.996889 |
|
Schaefer 2018 parcellation atlas |
Human |
Cortical parcellation atlas in MNI152 space; 100-1000 ROIs; 7 or 17 Yeo networks; 1mm and 2mm resolution |
atlas_schaefer_2018 |
Schaefer A, Kong R, Gordon EM, Laumann TO, Zuo XN, Holmes AJ, Eickhoff SB, Yeo BTT. Local-Global Parcellation of the Human Cerebral Cortex from Intrinsic Functional Connectivity MRI. Cereb Cortex, 29:3095-3114, 2018. |
|
SCIL b0 |
b0 |
GE (1.5, 3 T), Philips (3 T); Siemens (1.5, 3 T) |
scil_b0 |
<a href=’https://scil.dinf.usherbrooke.ca’>Sherbrooke Connectivity Imaging Lab (SCIL)</a> |
|
Sherbrooke 3 shells |
Human |
Multi shell data: b-vals: [0, 1000, 2000; 3500] (s/mm^2); 193 gradient directions |
sherbrooke_3shell |
<a href=’https://scil.dinf.usherbrooke.ca’>Sherbrooke Connectivity Imaging Lab (SCIL)</a> |
|
SNAIL dataset |
2 subjects: T1; Fractional Anisotropy (FA); 27 bundles |
bundles_2_subjects |
|||
Stanford HARDI |
Human |
HARDI-like multi-shell data: b-vals: [0, 2000] (s/mm^2); 160 gradient directions |
GE Discovery MR750 |
stanford_hardi |
<a href=’https://purl.stanford.edu/ng782rw8378’>Human brain diffusion-weighted MRI, collected with high diffusion-weighting angular resolution and repeated measurements at multiple diffusion-weighting strengths</a>. Rokem, A., Yeatman, J.D., Pestilli, F., Kay, K.N., Mezer A., van der Walt, S., and Wandell, B.A. (2015) Evaluating the Accuracy of Diffusion MRI Models in White Matter. PLoS ONE 10(4): e0123272. doi:10.1371/journal.pone.0123272 |
Stanford labels |
Human |
Gray matter region labels |
GE Discovery MR750 |
stanford_labels |
<a href=’https://purl.stanford.edu/ng782rw8378’>Human brain diffusion-weighted MRI, collected with high diffusion-weighting angular resolution and repeated measurements at multiple diffusion-weighting strengths</a>. Rokem, A., Yeatman, J.D., Pestilli, F., Kay, K.N., Mezer A., van der Walt, S., and Wandell, B.A. (2015) Evaluating the Accuracy of Diffusion MRI Models in White Matter. PLoS ONE 10(4): e0123272. doi:10.1371/journal.pone.0123272 |
Stanford PVE maps |
Human |
Partial Volume Effects (PVE) maps: Gray matter (GM), White matter (WM); Cerebrospinal Fluid (CSF) |
GE Discovery MR750 |
fetch_stanford_pve_maps |
<a href=’https://purl.stanford.edu/ng782rw8378’>Human brain diffusion-weighted MRI, collected with high diffusion-weighting angular resolution and repeated measurements at multiple diffusion-weighting strengths</a>. Rokem, A., Yeatman, J.D., Pestilli, F., Kay, K.N., Mezer A., van der Walt, S., and Wandell, B.A. (2015) Evaluating the Accuracy of Diffusion MRI Models in White Matter. PLoS ONE 10(4): e0123272. doi:10.1371/journal.pone.0123272 |
Stanford T1 |
Human |
T1 |
GE Discovery MR750 |
stanford_t1 |
<a href=’https://purl.stanford.edu/ng782rw8378’>Human brain diffusion-weighted MRI, collected with high diffusion-weighting angular resolution and repeated measurements at multiple diffusion-weighting strengths</a>. Rokem, A., Yeatman, J.D., Pestilli, F., Kay, K.N., Mezer A., van der Walt, S., and Wandell, B.A. (2015) Evaluating the Accuracy of Diffusion MRI Models in White Matter. PLoS ONE 10(4): e0123272. doi:10.1371/journal.pone.0123272 |
SyN data |
Human |
T1; b0 |
syn_data |
||
Taiwan NTU DSI |
DSI-like data; Multi-shell data: b-vals: [0, 308 ,615, 923, 1231, 1538, 1538, 1846, 1846, 2462, 2769, 3077, 3385, 3692, 4000] (s/mm^2); 203 gradient directions |
Siemens Trio |
taiwan_ntu_dsi |
National Taiwan University (NTU) Hospital Advanced Biomedical MRI Lab DSI MRI data |
|
Tissue data |
Human |
T1; denoised T1; Power map |
tissue_data |