Skip to Main Content
In neuroanatomical studies, the specimens are generally cut into serial sections that are processed to reveal the elements of interest. The third dimension lost during sectioning can be recovered by reconstructing three-dimensional graphical models of the studied structures. To reach statistical significance and to compare results from distinct experiments, data from different models must be combined into common representations. Due to biological and experimental variability, this requires a non-linear spatial normalisation step. In this paper, an algorithm is presented to normalise and map data into average models. The usefulness of the approach for elucidating spatial organisations in the nervous system is illustrated on rat neuroanatomical data.