Structural Informatics Group (SIG) logo
Home | Projects | Demos | Downloads | Publications | Local Info | About Us
Publications

Explaining the Brain to a Computer

Rosse, Cornelius and Tuttle, M. S. (2001) Explaining the Brain to a Computer. In Proceedings, Human Brain Project Annual Meeting.

Full text of this item is not available.

Abstract

The web-based resources required for supporting the mission of the Human Brain Project call for the development of "intelligent" tools and applications with an inherent capability for inference (reasoning). Formal representation of knowledge (concepts and relationships that exist between them) is a prerequisite for such computer empowerment. A feasible and promising first step toward this long-term goal is the establishment of controlled terminologies (or ontologies) within HPB's information domain. The principled process of generating such terminologies defines knowledge elements in "computer-understandable" format, which remains understandable also to humans. Moreover, such symbolic modeling is enlightening in that it improves understanding by uncovering the ambiguities and inconsistencies prevalent in traditional knowledge sources. Since normal and disordered functions may be conceptualized as attributes of anatomical structures ranging in size from the macromolecular to the macroscopic, we argue that a logical and comprehensive symbolic model of neuroanatomy can provide the foundation for the scalable and formal representation of functional as well as structural knowledge in HBP's information domain. We will illustrate the challenges entailed in symbolic modeling of neuroanatomy by comparing current controlled medical terminologies and the Digital Anatomist Foundational Model (FM), which is being developed as the anatomical enhancement of the Unified Medical Language System. The FM is a conceptualization of the physical organization (structure) of the human body without the brain and the spinal cord. We will explore the extent to which the principles, rules and definitions that guide the authoring of the FM could be applicable to neuroanatomy. Our intent is to illustrate the kinds of knowledge and their granularity that can be captured by principled modeling. A neuroanatomy ontology, with or without integration with the FM, should serve as a reference terminology, a scalable resource that can support interoperability between computers in the short term, and in the long term promote the evolution of knowledge-based tools and applications in neuroinformatics.

Item Type:Conference Paper
Keywords:brain mapping sig
Projects:Foundational Model of Anatomy, Human Brain Project
Subjects:SIG Publications
ID Code:100
Deposited By:Hinshaw, Kevin
Deposited On:01 May 2003
Alternative Locations:http://www.nimh.nih.gov/neuroinformatics/rosse2001.cfm