The UNC-Wisconsin Rhesus Macaque Neurodevelopment Database: A Structural MRI and DTI Database of Early Postnatal Development

Abstract

Rhesus macaques are commonly used as a translational animal model in neuroimaging and neurodevelopmental research. In this report, we present longitudinal data from both structural and diffusion MRI images generated on a cohort of 34 typically developing monkeys from 2 weeks to 36 months of age. All images have been manually skull stripped and are being made freely available via an online repository for use by the research community.

Keywords: brain development; computational atlases; diffusion tensor imaging; macaque; magnetic resonance imaging; neuroimaging; non-human primate.

Figure 1

This comparison of DTI scans for humans and rhesus macaques demonstrates that diffusion anisotropy is observable before a clear distinction between the white and gray matter is visible in the T1 weighted image. This difference in clarity and delineation is apparent in both the human and macaque images.

Figure 2

Illustration of actual scanning schedule used in this analysis. Each data point represents a scan of either a male (circle) or female (cross) subject. The data points are color coded based on the weight in kilograms of each subject at the time of the scan. Most subjects were scanned at five different times to create the data for this study.

Figure 3

Structural images were processed first by performing bias field correction, followed by rigid alignment to the atlas. Additionally, automatic and manual skull stripping was done. The DTI images were corrected for motion and eddy current artifacts, tensors were calculated, and finally, the T2 weighted image was registered to the baseline image and the brain mask was applied to the diffusion data to complete skull stripping.

Figure 4

A representative sample of the serial images acquired for a single subject in the database (Subject 032). Each subject was scanned at five age points. At each time point, a T1 and T2 scan was obtained and the image above shows axial, coronal, and sagittal views for each scan sequence.

Figure 5

These T1 images illustrate the difference between the automatic segmentation (top) and the manual corrected results (bottom) when performed on scans of a 2 week infant monkey. Automatic segmentations on younger subjects required significantly more manual corrections in part due to poor tissue contrast.

Figure 6

These images are an illustrative example the DTI results calculated from scans from all subjects. Property maps were generated at each scanning time point.

Figure 7

All data from this study is being made publically available and can be accessed from the homepage shown above. The data can be downloaded in whole or in part using a standard SFTP client. This repository will continue to be updated as we process the data.