Abstract: Objective To construct a voxel-based probabilistic diffusion tensor image (DTI) brain atlas in a standard space using medical image processing techniques in order to provide basic data for the white matter function research and the diagnosis and treatment of related diseases.Methods DTI data from 50 normal volunteers were acquired after excluding the existence of neurological diseases. The DICOM data were converted into the needed data format using one of the procedures in MRIcron software package. DtiStudio, DiffeoMap and AIR software were used for pre-processing, tensor calculation and normalization of the images. Average calculation was performed and the atlas was constructed in MATLAB software platform. Results A general DTI brain atlas was successfully developed from the 50 healthy adults and the maps were very clear. The white matter and fiber orientation were seen clearly in FA images and in the color map with specific colors representing different fiber orientations. BR>Conclusion Imaging processing techniques can be used to construct Chinese DTI probabilistic brain atlas. The maps are saved using a common data format and can be read by most medical image processing softwares and analysis softwares, thus may be used for the research and analysis of white matter function and related diseases conveniently. BR>

Key words: Atlas, Diffusion tensor image, Digital anatomy, Brain white matter, Human