Exploration and virtual experimentation in a local neuroscience database

Abstract

Work is currently being done by a number of groups to investigate the possibility of creating one or more neuroscience databases. These databases could provide: (1) a means for conveying complete descriptions of experiments; (2) a platform for virtual experiments; (3) and an interface where modeling and experimental results could be exchanged. This paper describes work towards creating a local database with these capabilities by creating a set of experimental neurophysiology software tools that are tied to a database of experiment descriptions via external scripting mechanisms. These tools provide a means of exploring results and data that can lead to insights into the data that suggest new avenues of research. Because the scriptability of the tools allows for automation of re-analysis of the data, virtual experiments for testing new hypotheses over large datasets becomes possible. Such virtual experiments are particularly relevant to computational neuroscience, as they allow hypotheses generated by models to be immediately tested without further collection of experimental data. © 1995.