LOGOS: a computational framework for neuroinformatics research

Abstract

Neuroinformatics presents a great challenge to the computer science community. Quantities of data currently range up to multiple-petabyte levels. The data itself are diverse, including scalar vector (from 1 to 4 dimensions), volumetric (up to 4 dimensional spatio-temporal), topological, and symbolic, structured knowledge. Spatial scales range from Angstroms to meters, while temporal scales go from microseconds to decades. Base data vary greatly from individual to individual, and results computed can change with improvements in algorithms, data collection techniques, or underlying methods. The authors describe a system for managing, sharing, processing, and visualizing such data. Envisioned as a "researcher's associate", it will facilitate collaboration, interface between researchers and data, and perform bookkeeping associated with the complete scientific information life cycle, from collection, analysis, and publication to review of previous results and the start of a new cycle.