We propose a knowledge base that combines numeric and symbolic knowledge about sulco-gyral brain cortex. This knowledge base is implemented using Web technologies. It is intended to be easily reusable in various application contexts such as teaching, decision support in neurosurgery and sharing of neuroimaging data for research purposes. Our analysis shows that (1) a formal representation of taxonomy and mereotopology, and (2) use of identity criteria to represent symbolic concepts, are needed to serve those applications.