Category:

Guide

Duration: 34:27

45 views

0

Systems-level neuroscience lacks a formal theoretical structure, relying on argumentation based on experimental findings expressed in the primary literature. Theoretical models may typically be represented as summary diagrams in a paperΓÇÖs discussion. Within a subject as complex and multifaceted as neuroscience, this lack of formalization leads inevitably to problems of information overload for individual researchers as it is a significant challenge to manage and manipulate large volumes of information from a distributed resource such as the literature and scientistsΓÇÖ own individual records. We present ΓÇÿNeuroScholarΓÇÖ, a knowledge base desktop application that specifically targets literature- and laboratory-based information, providing a structured knowledge engineering approach for neuroscience. It provides a general object-oriented data model to encapsulate complex data into entities, and a graph-theoretical approach that represents relations between entities edges between nodes in a graph. The system has frameworks for unit testing, plugins (to embed external applications within NeuroScholar), proxies (to export NeuroScholarΓÇÖs knowledge management capabilities to external applications) and knowledge acquisition based on questionnaires. Specialized plugins include (a) an annotation mechanism for pdf files (built with Multivalent, a third party library), (b) an electronic laboratory notebook component, (c) an annotation mechanism for vector graphics and (d) NeuARt, a neuroanatomical data viewer based on standard atlases which can also use the proxy framework to act as a standalone neuroanatomical data management tool. The knowledge acquisition subsystem provides an easy way to link free-form document annotation with structured knowledge representations for specific types of experiment. We are applying the system directly in two systems-level neuroscience laboratories, one focused on neuroanatomy, the other on neuroendocrinology. It is anticipated that NeuroScholar may provide a platform for theoretical research in neuroscience by delivering knowledge engineering capabilities directly to experimental scientists to facilitate analysis and communication.