Healthy-control steady-state hypoxia-targeted BOLD MRI reference and normative model maps
Description
This dataset contains group-level healthy-control steady-state hypoxia-targeted BOLD reference and normative model maps used in the manuscript “Voxel-wise temporal decomposition of hypoxia-targeted BOLD MRI reveals distinct timing and steady-state response signatures in glioblastoma”. Maps were derived from 20 healthy controls acquired during controlled isocapnic hypoxia and are provided in MNI space at 1.5 mm isotropic resolution. The steady-state hypoxia-BOLD response is reported in %BOLD/mmHg. The package includes mean and floored standard-deviation maps for conventional atlas-based comparison, beta0 and beta1 maps and floored residual standard-deviation maps for G-adjusted normative modeling, a final validity mask, an n-valid map, metadata, quality-control summary and README. No individual-subject imaging data, patient imaging data, clinical metadata or identifiable information are included. The dataset is intended to support transparency, methodological reuse and exploratory comparison and should not be interpreted as an age-general clinical normative standard.
Files
Steps to reproduce
The maps were derived from 20 healthy controls who underwent 3-T BOLD MRI during controlled isocapnic hypoxia. For each subject, the hypoxia-BOLD response was temporally decomposed voxel-wise and the O2-normalized steady-state response was calculated during the local plateau phase after exclusion of transition periods. Individual steady-state maps were normalized to MNI space and summarized voxel-wise across healthy controls. The package contains mean and floored standard-deviation maps for conventional atlas-based comparison. A global response scalar was calculated for each healthy control and voxel-wise linear regression was used to model the steady-state response as a function of this scalar. The package, therefore, also includes beta0 and beta1 maps, a floored residual standard-deviation map, an n-valid map, and a final validity mask. Conventional atlas-based comparison can be performed as: z = (patient_map - mean_map) / sd_floored G-adjusted deviation mapping can be performed as: expected = beta0 + beta1 * G_patient z_Gadjusted = (patient_map - expected) / residual_sd_floored All maps are in MNI space at 1.5 mm isotropic resolution. Voxels outside the final atlas validity region are set to NaN. The steady-state hypoxia-BOLD response is reported in %BOLD/mmHg.
Institutions
- University Hospital of ZurichZurich, Zurich