Skip to content

unidesigner/pyconto

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

50 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Python Connectome Toolbox

Why yet another network / graph toolbox?

There are already quite a number of very efficient and well-designed libraries and packages available that implement different graph representations, algorithms and layouting schemes.

See the recent discussion about a future Python Graph API

Existing Graph Libraries (under active development)

But: The Idea

Harnessing the power of NetworkX, nitime and pyhdf among others to provide an efficient framework for connectomics, investigations in structure and function of complex systems in the domain of neuroscience.

Exposition of a class hierarchy including static and dynamic networks based on Traits, abstracting from actual implementation of the graph data structures (replacable depending on ones needs).

Creation of NetworkAnalyzerObjects (see nitime) for a pipelined network analysis approach with maximum flexibility for different types of networks, different algorithms for similar problems, implementations of algorithms in different languages, caching mechanism to store computationally intensive network measures on large graphs. Creation of easy-to-read network analysis pipelines is wanted.

Providing a general playground for exploration of network dynamics by combining graph and time series data structures, thereby allowing to create novel algorithm and statistical procedures to quantify and understand structure and function relationships. In macroscale Connectomics in particular, this might be helpful in combining structural (e.g. Diffusion MRI) and functional (e.g. fMRI, EEG, MEG) data. In microscale Connectomics, for example in combining neural networks and their spiking activity (or any other quantifiable change over time, such as Voltage-Sensitive Dye Imaging).

Nonlinear Dynamics of Networks

http://www.cscamm.umd.edu/programs/ntd10/

Date & Location: 5-9 April, 2010, in College Park, MD

Organizers: Michelle Girvan (UMD), Ed Ott (UMD), Raj Roy (UMD) and Eitan Tadmor (UMD)

Description: The interconnection of many dynamical units to form a complex system can lead to unexpected collective behavior. This dynamics depends upon both the individual characteristics of the participating units, as well as the topological character and properties of the network of their connections. This workshop will focus on gaining understanding of general principles and techniques of analysis that will be of broad use in the many applications where networked system dynamics is a significant issue. Another aim of the workshop will be to highlight particularly important examples of applications where the issue of network dynamics arises.

Understanding the dynamics of networked systems is becoming an increasingly important and essential component in many areas of science and technology. Examples include social networks, communication and computer networks, gene networks, networks of neurons, etc. Dynamics on such networks include such problems as synchronization of temporal behavior of units composing a network, robustness of function to network damage (either intended or unintended), etc. The dynamics of networks themselves (i.e., change of network topological structure with time) is also an essential issue in many cases. Examples of issues in this area include adaptive evolution of network topology, formation and growth of networks, etc.

It is intended that all of the above, as well as related issues, will be open for discussion at this workshop. The two overarching goals of the workshop will be

  • To contribute to the understanding of common, basic principles of network dynamics, and
  • To uncover useful general analysis techniques for the study of these systems