One of the challenges of modern neuroscience is integrating massive data of different types derived in the studies of individual animals. This task requires a common spatial reference provided by brain atlases. The first atlases were two-dimensional presentations of structural data. Since brain is not flat, recently 3D brain atlases (computer based 3D models of brain structures) started to appear, to provide scientists with easier navigation within non-standard anatomical planes, and to improve positioning of different types of data within the brain volume.

The 3D atlases available so far were created with expensive tools and with techniques which make it difficult to repeat and improve the result and to integrate new data. To facilitate reproducible research and data sharing in this field we proposed a convenient format to store and distribute 2D atlas delineations or other compatible data which we call a Common Atlas Format (CAF). We also developed open-source (free) software for reconstruction of three-dimensional brain structures from 2D atlas data called 3D Brain Atlas Reconstructor (3dBAR) which makes the whole reconstruction process simple and reproducible. 3dBAR uses open file formats which simplifies interoperability with other neuroinformatics tools, adding new data, and preparation of new brain atlases.