A Chinese multi-modal neuroimaging data release for increasing diversity of human brain mapping

Abstract

The big-data use is becoming a standard practice in the neuroimaging field through data-sharing initiatives. It is important for the community to realize that such open science effort must protect personal, especially facial information when raw neuroimaging data are shared. An ideal tool for the face anonymization should not disturb subsequent brain tissue extraction and further morphological measurements. Using the high-resolution head images from magnetic resonance imaging (MRI) of 215 healthy Chinese, we discovered and validated a template effect on the face anonymization. Improved facial anonymization was achieved when the Chinese head templates but not the Western templates were applied to obscure the faces of Chinese brain images. This finding has critical implications for international brain imaging data-sharing. To facilitate the further investigation of potential culture-related impacts on and increase diversity of data-sharing for the human brain mapping, we released the 215 Chinese multi-modal MRI data into a database for imaging Chinese young brains, namely'I See your Brains (ISYB)', to the public via the Science Data Bank ( https://doi.org/10.11922/sciencedb.00740 ).

Fig. 1

Quality assessments on the ISYB sMRI data using both bar (the left column) and scatter (the right column) plots of the following quality metrics: (a) Signal-to-Noise Ratio (SNR) (b) Foreground to Background Energy Ratio (FBER) (c) Contrast-to-Noise Ratio (CNR) (d) Entropy Focus Criteria (EFC) (e) Full-Width Half Maximum (FWHM).

Fig. 2

Quality assessments on the ISYB rfMRI data using both bar (the left column) and scatter (the right column) plots of the following quality metrics: (a) Signal-to-Noise Ratio (SNR) (b) Foreground to Background Energy Ratio (FBER) (c) Entropy Focus Criteria (EFC) (d) Full-Width Half Maximum (FWHM), (e) Mean Framewise Displacement (FD_MEAN), (f) Derivative of time series and root-mean-square VARiance over voxelS (DVARS), (g) Global Correlation (GCOR).

Fig. 3

The group mean maps of the ISYB multi-modal neuroimaging data on the fsLR_32k cortical surfaces. Given each vertex on the surface, the mean values were calculated across all the 215 participants for the 12 metrics based on the following three MRI modalities: (a) sMRI (CT, SA, GMV, LGI), (b) dMRI (FA-WM, MD-WM, FA-GM, MD-GM), (c) rfMRI and aMRI metrics (CBF, ALFF, ReHo and VMHC).

Fig. 4

The group variability maps of the ISYB multi-modal neuroimaging data on the fsLR_32k cortical surfaces. Given each vertex on the surface, the standard deviation values were calculated across all the 215 participants for the 12 metrics based on the following three MRI modalities: (a) sMRI (CT, SA, GMV, LGI), (b) dMRI (FA-WM, MD-WM, FA-GM, MD-GM), (c) rfMRI and aMRI metrics (CBF, ALFF, ReHo and VMHC).

Fig. 5

The analytic framework of the evaluation of template effects on face anonymization. (a) The face masking toolkit takes raw individual unmasked MRI data with the registration from the unmasked head image to the head template. (b) The head templates include the default MNI152 and other two Chinese (CN200 and Chinese2020) templates. (c) During the face masking, a face mask is generated and used for restoring the face removed by the face masking pipeline. (d) The normalized mutual information between the removed face and the original face is calculated.

Fig. 6

The replicable template effects on face anonymization. Normalized mutual information between the face image from original data and the face image removed by the anonymization are visualized as violin plots with paired lines under the three different template (CN200, MNI152, Chinese2020) conditions. This template effect is replicated across three datasets: (a) ISYB-1, (b) ISYB-2, (c) ISYB. *p<0.05, ****p<0.0001.