Quantitative MRI (qMRI)🔗

Release Data🔗

Quantitative MRI release data include file-based (raw and processed data files in modality-specific formats) data:
See the Data Structure Overview for a full explanation of these data types.

• Raw BIDS stored under subject- and session-specific anat/ folders
• Derivatives generated by the SyMRI & qMRI Postproc pipelines

hbcd/
|__ rawdata/ 
    |__ sub-{ID}/
        |__ ses-{V0X}/
            |__ anat/
                |__ sub-{ID}_ses-{V0X}_acq-sub-<mrsLocAx|mrsLocCor>_run-{X}_T2w.nii.gz 
                |__ sub-{ID}_ses-{V0X}_acq-sub-<mrsLocAx|mrsLocCor>_run-{X}_T2w.json
                |__ sub-{ID}_ses-{V0X}_run-{X}_inv-sub-<0|1|2|3|4>_QALAS.nii.gz
                |__ sub-{ID}_ses-{V0X}_run-{X}_inv-sub-<0|1|2|3|4>_QALAS.json

hbcd/
|__ derivatives/ 
    |__ symri/  # SyMRI Derivatives
    |   |__ sub-{ID}/
    |       |__ ses-{V0X}/
    |           |__ anat/
    |               |__ sub-{ID}_ses-{V0X}_acq-QALAS_T1w.nii.gz (+JSON)
    |               |__ sub-{ID}_ses-{V0X}_acq-QALAS_T2map.nii.gz (+JSON)
    |               |__ sub-{ID}_ses-{V0X}_acq-QALAS_T2w.nii.gz (+JSON)
    |               |__ sub-{ID}_ses-{V0X}_acq-QALAS_desc-SymriContainer.log
    | 
    |__ qmri_postproc/ # qMRI Post-Proc Derivatives
        |__ sub-{ID}/
            |__ ses-{V0X}/
                |__ anat/  
                    |__ sub-{ID}_ses-{V0X}_desc-AsegROIs_scalarstats.tsv (+JSON)
                    |__ sub-{ID}_ses-{V0X}_desc-BilateralAsegROIs_scalarstats.tsv (+JSON)
                    |__ sub-{ID}_ses-{V0X}_desc-RegistrationQCAid.png (+JSON)
                    |__ sub-{ID}_ses-{V0X}_space-QALAS_desc-aseg_dseg.nii.gz
                    |__ sub-{ID}_ses-{V0X}_space-T2w_desc-QALAS_T2map.nii.gz (+JSON)

Data Acquisition🔗

QALAS🔗

QALAS is a multi-contrast MRI sequence that produces five brain volumes using turbo-flash readouts with varying T1 and T2 weightings. These volumes are combined to estimate T1, T2, and proton density (PD) maps. The sequence starts with a T2-preparation pulse, which adds T2 weighting to the first volume. An inversion pulse follows, imparting T1 weighting to the next four volumes:

Figure 1. QALAS Pulse Sequence Diagram (Fujita et al., 2024)

Acquisition time: ~5 minutes (Siemens), ~4 minutes (GE/Philips).

B1+ Fieldmap🔗

The HBCD protocol also includes a short B1+ fieldmap acquisition (<1 minute) to calibrate flip angle measurements (which can vary spatially due to variations and inhomogeneities in the B1+ field) needed for accurate T1, T2, and PD estimation. Because the transmit B1+ field is spatially smooth and changes gradually, coarse resolutions are sufficient for these scans, allowing for fast acquisition times. The methods by vendor used for the HBCD protocol were:

• GE & Philips: Actual Flip Angle Imaging (AFI) (Yarnykh 2007)
• Siemens: Pre-saturation turbo-flash readout

Acquisition time: ~30–45 seconds across all scanner types (Siemens, GE, and Philips).

Data Processing🔗

See the folder tree above for a full list of SyMRI and qMRI PostProc derivatives included in the release.

Relaxation (T1/T2) & Proton Density (PD) Maps🔗

For the HBCD study, the MRI working group adopted 3D-QALAS (Kvernby et al. 2014), a time-efficient 3D method that combines interleaved Look-Locker acquisition with a T2-preparation pulse. This approach simultaneously estimates longitudinal (T1) and transverse (T2) relaxation times, as well as proton density (PD) maps, from a single scan, and has been validated across major MRI vendors (Fujita et al. 2019).

Conventional neuroimaging typically relies on qualitative relaxation time-weighted images, e.g, T1w and T2w images, which reflect relative differences in relaxation times, but are strongly influenced by sequence parameters, participant positioning, and scanner hardware. These dependencies complicate biological interpretation and hinder quantitative comparisons across participants, sessions, and sites. The issue is particularly acute in pediatric neuroimaging, where rapid changes in free water distribution, iron, and myelination alter image contrast as a function of age. By directly measuring relaxation properties, quantitative MRI overcomes many of these limitations and provides more reliable markers of brain tissue microstructure (Deoni 2010; Does 2018).

Synthetic T1w/T2w Images🔗

Using 3D-QALAS data with the Synthetic MRI (SyMRI) toolbox, we also generate synthetic T1w (Sy-T1w) and T2w (Sy-T2w) volumes. These are created by substituting quantitative estimates of T1 and T2 relaxation times back into the MR signal equation (Bloch equations) for each sequence. This enables flexible generation of different contrasts without acquiring separate scans.

References🔗