An Introduction to XML in the Neurosciences
Sharon Crook, Arizona State University

An Introduction to NeuroML
Fred Howell, University of Edinburgh

NeuroML for Model Specification in ChannelDB and GENESIS
Dave Beeman, University of Colorado, Boulder

MorphML: An XML Application for Neuronal Morphology Data
Sharon Crook, Arizona State University

Building 3-D Network Models with neuroConstruct
Padraig Gleeson, University College London

Workshop Discussion: Current Issues and Future Development





NeuroML
http://www.neuroml.org

ChannelDB
http://www.modelersworkspace.org/channeldb/ChannelDB.html

MorphML
http://www.morphml.org

neuroConstruct
http://www.physiol.ucl.ac.uk/research/silver_a/neuroConstruct/




David Beeman is Professor Adjunct of Electrical and Computer Engineering at the University of Colorado, Boulder where he is developing educational materials for computational neuroscience using the GENESIS simulator. Dr. Beeman also participates in GENESIS development and support. Previously, he spent 20 years at Harvey Mudd College engaged in undergraduate teaching and research in computational solid state physics after receiving his Ph.D. in theoretical solid state physics from UCLA in 1967.

Robert Cannon is based at the Institute for Adaptive and Neural Computation in the School of Informatics at the University of Edinburgh, and works with Mike Hasselmo of Boston University on software development and modeling for building spiking networks to drive rodent behavior. His main interests are in the transition between traditional software architecture and biological model architecture - what software can learn from biology, and how software systems can make biological modeling more accessible.

Sharon Crook obtained a PhD in Applied Mathematics from the University of Maryland and did her postdoctoral work at the Center for Computational Biology at Montana State University. She currently holds a joint appointment in the Department of Mathematics and Statistics and the School of Life Sciences at Arizona State University. She is also a faculty affiliate of the Center for Rehabilitation Neuroscience and Rehabilitation Engineering of the Arizona Biodesign Institute at Arizona State University. In her primary research area of mathematical neuroscience, she uses mathematical models, analysis, and computer simulations to examine the behavior of neurons and neural networks and their role in neural computation.

Padraig Gleeson graduated from University College Dublin with a Master's Degree in Mathematical Physics. He spent a number of years doing R&D in the IT/Telecom industry. He is currently based in Angus Silver's Laboratory of Synaptic Transmission and Information Processing in University College London where he is developing neuroConstruct, a software package for creating networks of biologically realistic neurons. This work has been funded by the Medical Research Council, initially as part of a software development contract, and currently as part of a PhD in Physiology focusing on modeling information processing in the cerebellum.

Fred Howell studied microelectronics then worked in industry designing computers for Fujitsu and Digital (which is now HP). He returned to academics to obtain a PhD in Parallel Computing and is currently based at the Institute for Adaptive and Neural Computation in the School of Informatics at the University of Edinburgh. His research has included the development of simjava and collaborative work with Erik de Schutter on cerebellar network models in PGENESIS. In 2002, he and Robert Cannon founded Axiope in order to develop software for getting neuroscience data on the web, but he escaped back to academia recently.



There was a commitment from Fred Howell and Robert Cannon to update the NeuroML website and use SourceForge to help with mailing lists, version control, user contributions, etc.

After much discussion, one of the few items of consensus was that it is worthwhile to create a common data format in NeuroML for many of the most widely-used model formalisms. Because many of us focus on anatomical, physiological, or modeling characteristics at different levels of scale (sub-cellular, cellular, network, systems level), the easiest way to develop NeuroML will be to create several domain-specific description languages (eg. morphml, channelml) and then link them through NeuroML. This will also help potential users understand the underlying schema. This approach will also make it easier to take advantage of existing XML applications. For example, BrainML has elements for describing experimental electrophysiological data that could be extended to incorporate descriptions for model-generated data. Other existing XML applications might be used for describing sub-cellular biochemical reactions (SBML) or other cellular processes (CellML). NeuroML will define logical connections among these independent domains as needed.

The following working groups were formed:

Core Technology
Fred Howell
Graham Cummins
Greg Hood
Padraig Gleeson
Tom Morse
Gully Burns
Joe Svitak

Morphology
Sharon Crook
Padraig Gleeson
Michael Hines
Graham Cummins
Tom Morse
Joe Svitak

Channels
Dave Beeman
Tom Morse
Padraig Gleeson
Michael Hines
Sharon Crook

Channel Distribution
Graham Cummins
 Michael Hines
Sharon Crook
Padraig Gleeson

Connections
Bruce McCormick
Brad Busse
Aravind Aluri
Graham Cummins