Data from: Self-calibrated B1+ and B0 Field Inhomogeneities Estimation (SAFE) in 3D Magnetic Resonance Fingerprinting
Data files
Jan 20, 2026 version files 11.68 GB
-
README.md
3.46 KB
-
subject_1.mat
587.19 MB
-
subject_10.mat
540.23 MB
-
subject_11.mat
606.08 MB
-
subject_12.mat
562 MB
-
subject_13.mat
605.02 MB
-
subject_14.mat
604.75 MB
-
subject_15.mat
605.01 MB
-
subject_16.mat
602.88 MB
-
subject_17.mat
599.83 MB
-
subject_18.mat
598.53 MB
-
subject_19.mat
601.21 MB
-
subject_2.mat
574.43 MB
-
subject_20.mat
600.91 MB
-
subject_3.mat
540.66 MB
-
subject_4.mat
576.32 MB
-
subject_5.mat
564.01 MB
-
subject_6.mat
591.76 MB
-
subject_7.mat
581.43 MB
-
subject_8.mat
575.50 MB
-
subject_9.mat
564.68 MB
Abstract
This dataset contains whole-brain 1mm isotropic quantitative MRI maps (T1, T2, PD), coil sensitivity maps, and paired 4mm isotropic field maps (B0, B1+) of 20 healthy adult volunteers. All data were acquired on 3T GE and Siemens scanners, with approval of the institutional review board, and informed consent was obtained. The tissue quantitative maps were acquired using 3D Spiral Projection Imaging MR Fingerprinting (MRF), while the field maps were acquired from calibration scans. Such a resource enables the community to utilize the SAFE method on their own specific MRF acquisition protocols without the need to obtain additional paired training data.
Data Organization
subjects/
├── subject_1.mat # Siemens Vida, 6-min scan
├── subject_2.mat # Siemens Vida, 6-min scan
├── ...
├── subject_10.mat # Siemens Vida, 6-min scan
├── subject_11.mat # GE Premier, 2-min scan
├── subject_12.mat # GE Premier, 2-min scan
├── ...
└── subject_20.mat # GE Premier, 2-min scan
Subject Groups
| Subject Range | Scanner | Scan Duration | SNR Level |
|---|---|---|---|
| 1-10 | Siemens Vida | 6 minutes | Higher |
| 11-20 | GE Premier | 2 minutes | Lower |
Data Contents
Each .mat file contains the following variables:
Quantitative Maps (1 mm isotropic)
| Variable | Description | Units |
|---|---|---|
T1 |
Longitudinal relaxation time map | ms |
T2 |
Transverse relaxation time map | ms |
PD |
Proton density map | a.u. |
Field Maps (4 mm isotropic, from calibration scans, co-registered)
| Variable | Description | Units |
|---|---|---|
B0 |
Main magnetic field inhomogeneity map | Hz |
B1 |
Transmit field (B1+) efficiency map | relative units |
Auxiliary Data
| Variable | Description |
|---|---|
coil_sensitivity |
Multi-channel coil sensitivity maps |
Acquisition Parameters
Common Parameters (All Subjects)
- Field Strength: 3T
- Sequence: 3D Spiral Projection Imaging MR Fingerprinting
- Resolution: 1 mm isotropic (T1, T2, PD), 4 mm isotropic (B0, B1+ from calibration scans)
- Coverage: Whole brain
Siemens Vida (Subjects 1-10)
- Scan Duration: 6 minutes
- Characteristics: Higher SNR due to lower undersampling rate
GE Premier (Subjects 11-20)
- Scan Duration: 2 minutes
- Characteristics: Lower SNR due to higher undersampling rate
Data Format
- File Format: MATLAB .mat
- Compatibility: Compatible with MATLAB and Python (scipy.io)
Loading Data
MATLAB
data = load('subject_1.mat');
T1_map = data.T1_map;
T2_map = data.T2_map;
PD_map = data.PD_map;
B0_map = data.B0_map;
B1_map = data.B1_map;
coil_sens = data.Coil_map;
Python
from scipy.io import loadmat
data = loadmat('subject_1.mat')
T1_map = data['T1_map']
T2_map = data['T2_map']
PD_map = data['PD_map']
B0_map = data['B0_map']
B1_map = data['B1_map']
coil_sens = data['Coil_map']
Notes
- Field maps (B0, B1+) are acquired from separate calibration scans at 4 mm isotropic resolution and co-registered to the quantitative maps
- Quantitative maps (T1, T2, PD) are derived from the same MRF acquisition at 1 mm isotropic resolution
- Coil sensitivity maps can be used for parallel imaging reconstruction or sensitivity-related corrections
Citation
If you use this dataset in your research, we request that you consider citing appropriately.
Human subjects data
Data were acquired under IRB approval with informed consent obtained from all participants for public sharing. All datasets have been de-identified as .mat files by removing any personally identifiable information.