Several studies have suggested that many major functional areas can be located with respect to the cortical sulci, allowing the construction of an anatomo-functional map of the human brain. However, the main difficulty when establishing such a map based on cerebral sulci comes from high inter-subject variability of the cortical topography which makes the definition of a common standard nomenclature to describe the cortical topography a very tedious task. Only the observation and the recovery of statistical parameters representing the normal anatomical variability allow for the definition of objective criteria for the recognition of these structures. In order to address this problem, we have implemented SEAL (Sulcal Extraction and Automatic Labelling). Our modelling method automatically labels the cortical topography of a particular subject using spatial prior on the distribution of the main cortical folds and may require a minumum human intervention to correct some labellings.