bulbnet-reproduction

This is a model of the mammalian olfactory bulb for the NEURON simulator. The model contains only mitral and granule cells.

A full description may be found in: Davison A.P., Feng J. and Brown D. (2003) Dendrodendritic inhibition and simulated odor responses in a detailed olfactory bulb network model. J. Neurophysiol. 90:1921-35

To reproduce the figures from this paper, install NEURON, change to the code directory, run:

  $ nrnivmodl

then for each figure run:

  $ python run.py figure_2

and similarly for the other figures.

Time required to run simulations

The following times are for running with 32 threads on a workstation with dual Intel Xeon E5-2640 processors (8 cores, 2.6 GHz, 20 MB cache), 128 GB 2400 MHz DDR4 ECC RAM, writing to a 400 GB SSD SAS 12 Gbit/s hard disk. The time to run with a single core was about 10 times as long.

figure_2   1m 19s
figure_3   9m 10s
figure_4   3m 46s
figure_5   3m 56s
figure_7   2m 32s
figure_8   5m 25s
figure_9   14m 58s
figure_10  1h 46m
figure_11  15m 50s
figure_12  10m 42s
figure_13  1h 24m
figure_14  34m 19s
figure_15  8m 46s
figure_16  25m 32s
figure_17  6m 23s

Dependencies

• NEURON (https://neuron.yale.edu/neuron/)
• Gnuplot (http://www.gnuplot.info/)
• Python (v3.6 or more recent)
• NumPy
• PyYAML
• Jinja2
• epstopdf (optional - for converting figure to PDF)
• ImageMagick (optional - for converting figures to PNG)

For more information, contact andrew.davison@cnrs.fr