Data from: Variations in regional white matter volumetry and microstructure during the post-adolescence period: a cross-sectional study of a cohort of 1,713 university students

Human brain white matter undergoes a protracted maturation that continues well into adulthood. Recent advances in diffusion-weighted imaging (DWI) methods allow detailed characterisations of the microstructural architecture of white matter, and they are increasingly utilised to study white matter changes during development and ageing. However, relatively little is known about the late maturational changes in the microstructural architecture of white matter during post-adolescence. Here we report on regional changes in white matter volume and microstructure in young adults undergoing university-level education. As part of the MRi-Share multi-modal brain MRI database, multi-shell, high angular resolution DWI data were acquired in a unique sample of 1,713 university students aged 18 to 26. We assessed the age and sex dependence of diffusion metrics derived from diffusion tensor imaging (DTI) and neurite orientation dispersion and density imaging (NODDI) in the white matter regions as defined itein the John Hopkins University (JHU) white matter labels atlas. We demonstrate that while regional white matter volume is relatively stable over the age range of our sample, the white matter microstructural properties show clear age-related variations. Globally, it is characterised by a robust increase in neurite density index (NDI), and to a lesser extent, orientation dispersion index (ODI). These changes are accompanied by a decrease in diffusivity. In contrast, there is minimal age-related variation in fractional anisotropy. There are regional variations in these microstructural changes: some tracts, most notably cingulum bundles, show a strong age-related increase in NDI coupled with decreases in radial and mean diffusivity, while others, mainly cortico-spinal projection tracts, primarily show an ODI increase and axial diffusivity decrease. These age-related variations are not different between males and females, but males show higher NDI and ODI and lower diffusivity than females across many tracts. These findings emphasize the complexity of changes in white matter structure occurring in this critical period of late maturation in early adulthood.The dataset can be used with the associated R script "MRiShare_Regional_WMproperties_FrSysNeurosci2021.R". DATA-SPECIFIC INFORMATION FOR: MRiShare_WM_IDPs_FrontSysNeurosci2021.csv 1. Number of variables: 258 (plus ID column) 2. Number of cases/rows: 1,713 3. Variable List: ID: row number Sex: F-females, M-males Age: age in years [JHU tract abbreviation]_volume (Columns 5-32): JHU regional volume based on SPM12 (Jacobean-modulated) in mm3 [JHU tract abbreviation]_FA (Columns 33-59): JHU regional mean FA [JHU tract abbreviation]_MD (Columns 60-86): JHU regional mean MD (in mm2/sec) [JHU tract abbreviation]_AD (Columns 87-113): JHU regional mean AD (in mm2/sec) [JHU tract abbreviation]_RD (Columns 114-140): JHU regional mean RD (in mm2/sec) [JHU tract abbreviation]_NDI (Columns 141-167): JHU regional mean NDI [JHU tract abbreviation]_ODI (Columns 168-194): JHU regional mean ODI [JHU tract abbreviation]_ISOVF (Columns 195-221): JHU regional mean IsoVF Columns 222-233: Quality Control (QC) metrics related to structural scans -structQC_FLAIRtoT1_coregCF: Coregistration cost function for FLAIR to T1w image. -structQC_[T1 or FLAIR]_GM_invSNR: Inverse of signal-to-noise (SNR) ratio within GM signals in T1w or FLAIR image (based on SPM12 segmentation). The higher value indicates worse SNR. -structQC_[T1 or FLAIR]_WM_invSNR: Inverse of SNR ratio within WM signals in T1w or FLAIR image (based on SPM12 segmentation). The higher value indicates worse SNR. -structQC_[T1 or FLAIR]_CSF_invSNR: Inverse of SNR ratio within CSF signals in T1w or FLAIR image (based on SPM12 segmentation). The higher value indicates worse SNR. -structQC_[T1 or FLAIR]_[WMGM or GMWM]_invCNR: Inverse of WM-to-GM (for T1w) or GM-to-WM (for FLAIR) contrast-to-noise (CNR) ratio (based on SPM12 segmentation). Higher value indicates worse CNR. The higher value indicates worse CNR. -structQC_[T1 or FLAIR]_GMCSF_invCNR: Inverse of GM-to_CSF CNR ratio in T1w or FLAIR image (based on SPM12 segmentation). The higher value indicates worse CNR. Higher value indicates worse CNR. -structQC_FS6_invEuler: inverse of the Freesurfer's Euler number. The higher number indicate poor quality. Columns 234-251: QC metrics related to diffusion scans and DTI computation -DWIQC_[b0, b300, b1000 or b2000]_mean_afni_outlier_post: Mean proportion of outlier voxels based on AFNI 3dOutcount tool. -DWIQC_[b0, b300, b1000 or b2000]_inv_tSNR: Inverse of temporal SNR for each b value image. The higher value indicates lower tSNR. -DWIQC_[b300, b1000 or b2000]_inv_CNR: Inverse of CNR for each non-zero b value image, outputted by Eddy. The higher value indicates worse CNR. -DWIQC_shellNS_relative_rms: Mean relative RMS for the entire DWI scan computed using relative RMS outputted by Eddy. -DWIQC_shellNS_relative_restricted_rms: Mean relative restricted RMS for the entire DWI scan computed using relative restricted RMS outputted by Eddy. It measures the displacement caused by subject motion. -DWIQC_shellNS_eddy_mean_outlier_slices: Mean number of outlier slices per volume computed using the number of outlier slices outputted by Eddy. -DWIQC_shellNS_eddy_mean_outlier_[stdev or sqr_stdev]: Mean values of standard deviations off the mean (stdev) or the square root of the mean squared (sqr_stdev) difference between the observation and prediction, outputted by Eddy. -DTIQC_mean_residuals: Mean residuals inside the brain mask when fitting DTI. -DTIQC_percent_implausible_voxels: Proportion of physically implausible voxels (cf. Tournier et al, 2011) Columns 252-259: QC metric related to spatial normalisation. -NormQC_[wm, FA, MD, AD, RD, ICVF, ODI, or ISOVF]_inv_r2z: Inverse of the Fisher Z-transformed Person's correlation r measuring the similarity of the individual WM volume, DTI/NODDI scaler maps in stereotaxic space to the cohort average map. The higher value indicates more deviation from the average map. 4. Missing data codes: Four subjects did not have any volumetric or DTI/NODDI values for fornix (FX), as the WM probability map did not overlap with this small ROI in the standard space. For the same reason, one subject was missing data for the tapetum (TAP). In addition, for corticospinal tract (CST; n = 4) and inferior cerebellar peduncle (ICP; n = 6) ROIs, mean DTI/NODDI values were not computed in the pipeline since these ROIs extended beyond the bounding box of the DWI-derived images in the standard space. DATA-SPECIFIC INFORMATION FOR: JHI_27ROI_info.csv 1. Number of variables: 6 2. Number of cases/rows: 27 3. Variable List: tract_name: Long name of the JHU ROI used in the associated publication. abb_tract_name: Abbreviated JHU ROI names used in the associated publication as well as column names in the MRiShare_WM_IDPs_FrontSysNeurosci2021.csv hemi: "Both" medially located ROIs and "RL" ROIs present in each hemisphere size_group: Arbitrary grouping based on the ROI volume of the JHU atlas mori_group: One of the four groups based on Mori et al. (2008) that describes the ROIs in the JHU atlas. "Brainstem", "Projection", "Association", or "Commissural". atlas_vol_hemi_averaged: Label volume in the atlas, averaged between the two hemispheres for "RL" ROIs that exist ion each hemisphere. 4. Missing data codes: No missing data Funding provided by: Agence Nationale de la RechercheCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100001665Award Number: ANR-10-LABX-57; ANR-16-LCV2-0006; ANR-10-COHO-05-01Funding provided by: Horizon 2020*Crossref Funder Registry ID: Award Number: No 640643Funding provided by: Conseil Régional AquitaineCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100009468Award Number: 4370420Funding provided by: Fondation pour la Recherche MédicaleCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100002915Award Number: DIC202161236446Funding provided by: Horizon 2020Crossref Funder Registry ID: http://dx.doi.org/10.13039/501100007601Award Number: No 640643The dataset is based on magnetic resonance imaging (MRI) data collected as part of the MRiShare database. The anatomical and diffusion-weighted imaging data from those under 26 years of age from the MRiShare database were processed as described in the associated publication. The dataset contains regional white matter (WM) phenotypes based on JHU white matter label atlas as described in the paper and includes WM volume, mean DTI and NODDI metrics in each region defined in the JHU atlas. It also contains quality control (QC) metrics of the anatomical and diffusion-weighted imaging data