Rick Herrick, lead developer of the XNAT team, recently co-authored a paper for Frontiers in Neuroinformatics titled “Data dictionary services in XNAT and the Human Connectome Project.” The concept of the Data Dictionary, as a way of codifying distributing user-friendly definitions of data types and attributes, was a necessary development for the Human Connectome Project and its ConnectomeDB platform.
From the paper’s abstract:
The data dictionary service provides a framework to define key relationships between data elements and structures across the XNAT installation. This includes not just core data representing medical imaging data or subject or patient evaluations, but also taxonomical structures, security relationships, subject groups, and research protocols. The data dictionary allows users to define metadata for data structures and their properties, such as value types (e.g., textual, integers, floats) and valid value templates, ranges, or field lists.
For a user-readable example of a data dictionary, here is the set of open-access data categories, instruments and attributes in use for the Human Connectome Project: