Informatics

There is no greater challenge to the biomedical information processing community than applying computer technology to research and collaboration in neuroscience: neuroinformatics. Quantities of data range up to multiple-petabyte levels. The information itself is extraordinarily diverse, including scalar, vector (from 1 to 4 dimensions), volumetric (up to 4 dimensional spatio-temporal), topological, and symbolic, structured domain knowledge. Spatial scales range from Angstroms to meters, while at the same time temporal scales go from microseconds to decades. Base information varies greatly from individual to individual, while the results computed from this data and domain knowledge derived from such computation can change with improvements in algorithms, data collection techniques, or the underlying scientific methods. Coupled to this data complexity is the peculiarities of human information absorption, processing, and interaction.

The BCL is developing LOGOS: a system for storing, sharing, processing, visualizing, and reasoning about neuroscience information. This system is envisioned as a ``researcher's associate'', facilitating collaboration among researchers, serving as an interface between researchers and the data they collect and analyze, and taking care of all of the data and knowledge management associated with the complete scientific information life cycle, from experiment design, simulation, collection, processing, analysis, visualization, inference, generalization, and publication to review of previous results and the beginning of a new cycle.

Scientific Data/Knowledge Management

The heart of LOGOS is a data/knowledge based system described in a white paper and in an older publication (Stiber et al., 1997) and by a student project (Goebel & Mager, 2000).

Simulation

Two types of custom simulation environments have been developed within the BCL:

NeuronPC/XNeuron/CNeuron/MATLAB_Neuron
A family of simulators targeted at detailed neuromotor simulations which include a number of nonlinear dynamics analysis tools. These simulators include versions with graphical users interfaces written in C for PCs and Unix workstations, as well as a GUI simulator written in C and MATLAB. There is also a C version with no user interface, meant to be incorporated into shell scripts for automating large number of simulations.

Currently, these simulators implement models of the crayfish slowly adapting stretch receptor organ (SAO) and its associate prototypical inhibitory synapse, inhibitory long-term potentiation (LTP) rules for synapse modification, and muscle fibers.

CRESSA
A leaky integrator model which incorporates inputs with cluster process statistics. This has been used to investigate the effects of presynaptic correlation among multiple synapses on postsynaptic behavior.

Data Analysis and Capture

Under construction