Search

Methods for imaging the architecture of the brain at high resolutions, and across large volumes, are rapidly improving. With the convergence of high-resolution datasets and new computational approaches for processing them, fully and semi-automated methods for studying the brain will soon be within reach. However, there are many challenges in developing data-driven strategies for brain mapping with images at cellular and sub-cellular resolutions. This review highlights some key challenges in building models of brain structure from imaging datasets; we describe some existing efforts to tackle these challenges and potential solutions moving forward. Finally, we discuss the need for concerted community efforts to adopt common standards and coordinate systems for brain mapping, which will enable us to achieve robust and scalable solutions that work across different brain models and can accommodate the intrinsic variability both between and within high-resolution neuroimaging datasets.