The availability of high resolution MRI has made it possible to construct detailed atlases of the brains of live animals for the first time. We have been using 16.4T scans of mouse brain and 9T scans of rat brain to create detailed maps that rival those currently presented in the histological atlases published by our laboratory. In the preparation of these MRI atlases we have found that delineation of structures is enhanced with the use of images that are an average of scans from many animals (18 in the case of our mouse studies). This results in a far better signal to noise ratio. Furthermore, we have had access to a range of different sequences in analyzing the data that has been acquired, in order to emphasize different contrasts in the tissue. These include the use of different relaxation times (T1 and T2), differences in proton density, and diffusion tensor imaging (DTI). In our rat studies, the DTI image processing allowed the production of six different DTI data sets (ADC, DWI, LD, RD, FA, and a directionally encoded FA color map). Each of these image sets revealed different anatomical features. The use of this wide range of contrast sequences is somewhat comparable to the use of different marker stains in histological material. Using these techniques we are able to confidently identify over 500 structures in a rat brain, compared to over 1000 in our histological atlas. However, MRI atlases cannot match histological atlases in a number of respects, such as access to data from a wide variety of specific antibody markers, and can yield very high resolution data on neuronal anatomy. We anticipate that MRI atlases will deservedly gain currency, but will still be used alongside histological atlases for many years to come.