dHCP Fourth Data Release (April 2024)

Reconstruction pipeline

The reconstruction pipeline was developed by the team at King’s College London.

Motion corrected volumetric reconstructions of multi-slice inversion recovery T1- and turbo spin echo T2-weighted images are obtained by extending the aligned sensitivity encoding (SENSE) method¹ to the multi-slice case². Corrections are performed for both within-plane and through-plane motion from partial k-space information. The two acquired orthogonal stacks are integrated by a super-resolution scheme³. Methods and example data for the aligned SENSE method are available at https://github.com/mriphysics/multiSliceAlignedSENSE/releases/tag/1.0.1.

Inputs and outputs

Path: rawdata/sub-{subid}/ses-{sesid}

The reconstruction pipeline generates the files listed in the Reconstruction pipeline section of the directory structure summary.

Standard magnitude and phase reconstructions are provided for all native acquired anatomical image stacks, with the different orientations and repeat acquisitions labelled by sequence run number as follows:

Motion corrected and super-resolved reconstructions, and phase images, are included for every acquired multislice stack:

Primary anatomical outputs

The final motion corrected slice-to-volume reconstructed T1w and T2w volumes are:

An additional T1 3D MPRAGE volume is also provided, but currently not used in processing pipelines:

For fMRI and dMRI, simultaneous multi-slice (SMS) echo planar imaging (EPI) is reconstructed using the extended SENSE technique⁴, with details described elsewhere^5,6,7; sensitivity estimates from a conventional reference scan are refined with the information from non-SMS reference acquisitions with matched readouts to promote matched coil map and image distortions. As for dMRI, complex data retrieval is performed by the generalized singular value shrinkage (GSVS) denoising technique using noise measures performed during the acquisition⁸. Methods and example data for the GSVS method are available at https://github.com/mriphysics/complexSVDShrinkageDWI/releases/tag/1.1.0.

Primary dMRI outputs

dMRI reconstructions are provided with and without denoising, along with phase images, and additional chi^2 maps from the reconstruction required for the dMRI (SHARD) pipeline. A third reconstruction used by the dMRI (EDDY) pipeline is included, this matches the version from the 2nd data release, the main difference being that residual fat artefacts are suppressed in the reconstruction pipeline:

Primary outputs for fMRI

There is one primary reconstruction for the resting state fMRI data, along with single-band reference scans (typically one before and one after resting state run), and a spin-echo EPI with matched readout for field estimation. In addition, all phase images are provided.

Primary outputs for field mapping

Data from calibration scans are provided for B1+ and B0 field estimates, using the DREAM and dual-gradient echo methods, respectively. Reconstructed magnitude and phase images are provided, along with the following calculated field maps as the primary outputs:

The full output list can be found in the Reconstruction pipeline section of the directory structure summary.

References

1. Cordero-Grande, L., Teixeira. R. P. A. G., Hughes, E. J., Hutter, J., Price, A. N., and Hajnal, J. V. Sensitivity encoding for aligned multishot magnetic resonance reconstruction IEEE Transactions on Computational Imaging (2016), 2(3): 266-280. DOI: 10.1109/TCI.2016.2557069

2. Cordero-Grande, L., Hughes, E. J., Hutter, J., Hutter, J., Price, A. N., and Hajnal, J. V. Three-Dimensional Motion Corrected Sensitivity Encoding Reconstruction for Multi-Shot Multi-Slice MRI: Application to Neonatal Brain Imaging Magnetic Resonance in Medicine 2018, 79(3): 1365-1376. DOI: 10.1002/mrm.26796

3. Kuklikova-Murgasova, M., Quaghebeur, G., Rutherford, M. A., Hajnal, J. V., and Schnabel, J. A. Reconstruction of fetal brain MRI with intensity matching and complete outlier removal Medical Image Analysis (2012), 16(8): 1550-1564. DOI: 10.1016/j.media.2012.07.004

4. Zhu, K., Dougherty, R. F., Wu, H., Middione, M. J., Takahashi, A. M, Zhang, T., Pauly, J. M., Kerr, A. B. Hybrid-space SENSE reconstruction for simultaneous multi-slice MRI IEEE Transactions on Medical Imaging, 35(8) (2016):1824-1836. DOI: 10.1109/TMI.2016.2531635

5. Cordero-Grande, L., Price, A. N., and Hajnal, J. V. Comprehensive CG-SENSE reconstruction of SMS-EPI ISMRM 2016: 3239.

6. Cordero-Grande, L., Hutter, J., Price, A., Hughes, E., and Hajnal, J. V. Goodness of fit factor in SENSE reconstruction: a tool for pseudolesion detection and fat unfolding ESMRMB 2016: 458.

7. Hennel, F.,Buehrer, M., von Deuster, C., Seuven, A., and Pruessmann, K. P. SENSE reconstruction for multiband EPI including slice-dependent N/2 ghost correction Magnetic Resonance in Medicine (2016), 76(3): 873-879. DOI: 10.1002/mrm.25915

8. Cordero-Grande, L., Christiaens, D., Hutter, J., Price, A. N., and Hajnal, J. V. Complex diffusion-weighted image estimation via matrix recovery under general noise models Neuroimage (2019), 200: 391-404. DOI: 10.1016/j.neuroimage.2019.06.039