CNS 2011 workshop: Emerging standards for network modeling in neuroscience
The growing number of large-scale neuronal network models has created a need for simulator-independent description languages to ease model sharing and facilitate the replication of results across different simulators. To coordinate and promote community efforts towards such standards, the International Neuroinformatics Coordinating Facility (INCF) has formed its Multiscale Modeling program, and has assembled a task force to propose a declarative computer language for descriptions of large-scale neuronal networks: NineML. The INCF Task Force responsible for NineML development contains delegates from several relevant projects like the Blue Brain Project, GENESIS, KInNeSS, MOOSE, NEST, NeuroML, NEURON and PyNN.
This workshop aims to present an overview of the tools available to the community that support the development and exchange of network models. One key new community development is NineML, which provides a simulator independent description of model dynamics. At the same time, the next generation of NeuroML, (NeuroML V2.0) is under active development and will introduce important new advances in interoperable model description.
The focus of the morning session will be to present the current development in model description languages, with the presentation of two complementary initiatives, NineML and NeuroML. This will be followed by a presentation on the Connection Set Algebra, a recent development for describing network connectivity, which is currently being integrated with NEST and PyNN.
The afternoon session will then be focused on demonstrating interoperability using model description, APIs and runtime data exchange, with use cases that demonstrate the integration of these elements. We will present the current status of the interactions between NeuroML/LEMS and NineML, the integration of NineML within PyNN and the interrelation between PyNN and the multi-simulation coordination software MUSIC, developed as an INCF initiative. We will conclude this session by a presentation of the potential use of NineML as a native language for multiscale modeling. As all the presented initiatives were developed for and rely on the comments of the computational neuroscience community, we would like to invite the workshop participants to share their experiences, their questions and comments to help us to further develop these tools and their interactions for the community of modelers.
Finally, deeply aware of the dynamism of the community, we would like to invite interested scientists to present in a lightning talk (4 + 1 minutes) their related work, developed outside of the major initiatives. We will select 3 to 4 of these talks to conclude this workshop. To participate, email your talk idea to Malin Sandström (firstname.lastname@example.org) latest Friday July 8.
The target audience is computational neuroscientists, developers of modeling tools, and anyone interested in replication and reuse of modeling results across platforms and research groups.
Lightning talk speakers
Welcome and Introduction - Sean Hill
What is NineML: Motivation and Concepts
-Sean Hill (30+10 minutes)
Introduce NineML 1.0 (Synapses and Neurons - Abstraction and User Layer) as new way to describe model dynamics in a simulator independent manner
NeuroML v2.0: Status and future directions
-Sharon Crook (30+10 minutes)
Connection Set Algebra: introduction and use cases
–Mikael Djurfeldt (30+10 minutes)
Proposed standardized connection generation interface
At the heart of interoperability: Use case presentations
LEMS, NeuroML V2.0 and NineML
-Padraig Gleeson/Robert Cannon (30 + 10 min)
Interoperability between NineML, NeuroML and LEMS: Use cases
-Michael Hull (30+10 min)
Presentation of the Python lib9ml support library
-Andrew Davison (30+10 min)
Presentation of PyNN and Use cases with NineML, NeuroML, MUSIC and NEURON, NEST, PCSIM, Brian
MUSIC - an update
-Mikael Djurfeldt (30 + 10 min)
Presentation of multi-simulation using MUSIC and some recent developments
Workflow for Network Modeling with NineML
-Ivan Raikov (30 +10 min )
Opening up to other initiatives
-3-4 Lightning talks (4 + 1min)
Multiscale Modeling and Interoperability using GENESIS 3
- Dave Beeman (4+1 min); GENESIS 3
Modigliani, a fast, stochastic simulation framework for heterogeneous multiscale neuronal modelling
- Ali Neishabouri (4+1 min); Faisal lab
ClamVis: a user interface for MIIND (Multiple Interacting Instances of Neuronal Dynamics )
- Dave Harrison (4+1 min)
Discussion - Future directions, needs, opportunities (30 minutes)