Debian Codegraph @

docs/autodocs/source/index.rst

raw
1 .. Brain Modeling Toolkit documentation master file, created by
2 sphinx-quickstart on Tue Sep 26 10:33:33 2017.
3 You can adapt this file completely to your liking, but it should at least
4 contain the root `toctree` directive.
5
6
7 Welcome to the Brain Modeling Toolkit
8 =====================================
9
10 .. toctree::
11 :maxdepth: 2
12 :titlesonly:
13 :hidden:
14
15 Installation Guide <installation>
16 Building Networks <builder>
17 simulators
18 analyzer
19 tutorials
20 ad_tutorials
21 Example Networks <examples>
22 Github Profile <https://github.com/AllenInstitute/bmtk.git>
23 workshop
24 registration
25 contributors
26 how_to_cite
27 Source Documentation <bmtk/bmtk>
28
29
30
31 The Brain Modeling Toolkit (BMTK) is a python-based software package for building, simulating and analyzing large-scale
32 neural network models. It supports the building and simulation of models of varying levels-of-resolution; from
33 multi-compartment biophysically detailed networks, to point-neuron models, to filter-based models, and even
34 population-level firing rate models.
35
36
37 .. figure:: _static/images/levels_of_resolution_noml.png
38 :scale: 70%
39
40 The BMTK is not itself a simulator and will utilize existing simulators, like NEURON and NEST, to run different types of
41 models. What BMTK does provide:
42
43 * A unified inteferce across different simulators, so that modelers can work with and study their own network models
44 across different simulators without having to learn how to use each tool.
45 * An easy way to setup and initialize network simulations with little-to-no programming necessary
46 * Automatic integration of parallelization when running on HPC.
47 * Extra built-in features which the native simulators may not support out-of-the-box.
48
49 The BMTK was developed and is supported at the `Allen Institute for Brain Science <https://alleninstitute.org/>`_ and
50 released under a BSD 3-clause license. We encourage others to use the BMTK for their own research, and suggestions and
51 contributions to the BMTK are welcome.
52
53 The latest release, previous releases, and current development can be found at `<https://github.com/AllenInstitute/bmtk>`_
54
55
56 The BMTK Workflow and architecture
57 ----------------------------------
58 BMTK can readily scale to run models of single neurons, and even single compartments, and for all different types of
59 neuronal networks. However BMTK was designed for very-large, highly optimized mammalian cortical network models.
60 Generating the connectivity matrix could take hours or even days to complete. Plus once we had the base-line model we
61 wanted to run against a large variety of conditions and stimuli, so that we could directly compare with existing in-vivo
62 recordings (see Allen Brain Observatory).
63
64 .. figure:: _static/images/bmtk_workflow.jpg
65 :scale: 100%
66
67 Unlike other simulators, BMTK separates the process of building, simulating, and analyzing the results. First a fully
68 instantiated base-line version of the model is built and saved to a file so each time a simulation is ran it takes only
69 a small fraction of the time to instantiate the simulation. Results are also automatically saved to a disk. BMTK and
70 the format it uses (SONATA, see below) makes it easy to dynamically adjust cell and synaptic parameters so that multiple
71 iterations of a simulation be done as fast as possible
72
73 As such BMTK can be broken into three major components:
74
75 * The Network Builder [:py:mod:`bmtk.builder`] - Used to build network models
76 * The Simulation Engines [:py:mod:`bmtk.simulator`] - Interfaces for simulating the network
77 * Analysis and Visualization Tools [:py:mod:`bmtk.analyzer`] - Python functions for analyzing and visualizing the
78 network and simulation results
79
80 .. figure:: _static/images/bmtk_architecture.jpg
81 :scale: 45%
82
83 The components can be as one workflow or separately by themselves. BMTK utilizes the SONATA Data format (see next section)
84 to allow sharing of large-scale network data. As such, it is possible to use the Network Builder to build the model but
85 another tool to run the model. Or if another model has been built by someone else and saved in SONATA format BMTK will
86 be able to simulate it.
87
88
89 SONATA
90 ------
91 SONATA is a multi-institutional developed standardized, cross-platform data format for storing large scale networks and
92 simulation results. The BMTK utilizes SONATA when building and simulating networks, so much of what is being described
93 in the documentation and tutorials will be based on SONATA. For more information see the
94 `SONATA github page <https://github.com/AllenInstitute/sonata>`_.
95
96
97 VND
98 ---
99 `Visual Neuronal Dynamics <https://www.ks.uiuc.edu/Research/vnd/>`_ is a software package for 3D visualization of neuronal network models. VND can be used to check and inspect models, such as those created by BMTK, as well as to visualize the activity output. Images and movies made in VND can also be used to showcase or schematize models for presentations and publications. VND is developed in a collaboration between researchers at the University of Illinois at Urbana-Champaign and Allen Institute.
100
101
102 References
103 ----------
104 See the paper about BMTK: `link <https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1008386>`_.
105
106 For information on how to cite BMTK, please see: `link <how_to_cite.html>`_.
107
108
109
110
111 Acknowledgements
112 ----------------
113 We wish to thank the Allen Institute for Brain Science founder, Paul G. Allen, for their vision, encouragement, and support.