Parametric Anatomical Modeling is a method to translate large-scale anatomical data into spiking neural networks. PAM is implemented as a Blender addon.
Complex connection patterns between neurongroups are described by layers and a combination of simple mapping techniques between layers
2d layers define the location of neurons and their projection directions. Probability functions for pre- and post-synaptic neurons are applied on the surface of the synaptic layer to determine connections between two neuron groups.
A layer is defined as a 2d manifold (a deformed 2d surface). Each point on a layer is described by x, y, and z coordinates and u,v-coordinates which may correspond to anatomical axes.
In order to create axonal and dendritic connections in 3d space, neuron positions are mapped between layers. When the internal mesh-structure between layers is identical, neurons can be directly mapped using topological mapping. Otherwise, normal-, Euclidean- and random-based mapping are available.
- CVS-export of connection/distance matrix for external use
- Python data-import module for NEST Framework available at missing link.
-
Download the latest official release (a complete list can be found here) or get a snapshot of this GIT repository
-
Install the pam-folder as Add-On in Blender (how to install an Add-On). You can, for example, copy the pam-folder into the 2.X/scripts/addon-folder of your Blender version. If you downloaded the ZIP-file, you must only install the pam-folder as addon for Blender
-
In Blender, go to preferences - addons, search for PAM and activate it (optionally click on 'Save Preferences')
-
In the 3d view a new panel should appear on the left side, called "PAM". This panel contains function to load and save connectivity data and to visualize them.
The repository contains two examples, a very simple one and the hippocampal model, which could help to understand how to use PAM.
The Wiki contains a list of all mapping techniques and a description of the approach used to generate synapses.
Here is a list of videos that introduce PAM and show some features of it.
An importer of with PAM generated data to NEST can be found here
PAM is continuously under development. Any support and active participation in the development of PAM is very much welcome.
One long-term goal of our research is to structurally and functionally compare brain structures across different species using PAM. Therefore, we are looking for people who can assist us in creating the anatomical layout of brain structures and the projections of neurons across neural layers.
Dr. Martin Pyka Mercator Research Group "Structure of Memory", Ruhr-University Bochum
m.pyka at rub.de +49-234- 32 24682
Source code and documentation copyright (c) 2013-2015 Martin Pyka, Sebastian Klatt
pam is licensed under the GNU GPL v2 License (GPLv2). See license file for full text.