Code of a CA1 field model

This is a large-scale biologically plausible network model of rhythms in the hippocampal CA1 field.

Dependencies

Model is runned under Python 3.7. OS is Ubuntu 18.04 or Ubuntu 20.04, another os in not tested.

We will need to install git, pip, and mpich for MPI support:

sudo apt update
sudo apt install git python3-pip mpich

Python packages for simulations:

• NEURON 7.6+
• LFPsimpy with our modifications https://github.com/ivanmysin/LFPsimpy
• mpi4py

Packages for saving, processing and plotting:

• numpy, scipy, matplotlib
• h5py
• elephant

Installation pakages via pip:

sudo pip3 install numpy scipy matplotlib neuron mpi4py h5py elephant

Directories and files

 /cells - directory with models of cells (hoc files)
 /mod - directory with mechanisms for cells (mod files)
 /Results - directory for saving simulation results and plots
 base_model.py - main python file
 basic_parameters.py - python file containing all parameters of simulation
 nontheta_state_params.py - parameters for non-theta state
 presimulation_lib.py - helper functions for calculating some simulation parameters
 simulation_parallel.py - contains function *run_simulation*, it builds and runs the simulation.
 process.py - process hdf5 file
 processingLib.py - library of funtions, which are used for signals porcessing
 plot_result.py - library of funtions, which are used for plotting
 figure_n.py - files, which are plot figures for article. n is number of figure in article

How to run

You need:

• Clone this repository and repository with LFPsimpy https://github.com/ivanmysin/LFPsimpy to same directory. Open terminal in working directory and execute comants:

  git clone https://github.com/ivanmysin/CA1_rhythms_model
  git clone https://github.com/ivanmysin/LFPsimpy
  

• Compile .mod files in ./mods directory with nrnivmodl

  cd ./CA1_rhythms_model/mods 
  nrnivmodl
  cd ../
  

• Run in terminal:

    mpiexec -n nthreads python3 base_model.py
  

where nthreads is numbers of CPU cores used for simulation Simulation results are saved to hdf5 file. python3 is the default command for calling the python 3 interpreter, however, it may be different on your system. Substitute the call of the interpreter for which you have installed dependencies.

• To process the results, run process.py. It will process and save in the same file, wavelet spectra, bands, distribution of neurons by rhythm phases, etc.

  python3 process.py
  

• Run one of the figure_2(4, 5, 7, 8).py file to plot figure. For example:

  python3 figure_2.py
  

Full simulation need 70 Gb RAM! Full simulation: 9000 pyramidal cell and CA3 gererators, 200 PV basket cells etc. You can change cells numbers in basic_parameters.py file to reduce RAM usage and computational resources.

Structure of HDF5 file with results

You can use any free convenient program for viewing hdf5 files, for example, HDF COMPASS or hdfview.

Path for reading datasets from hdf5 file:

simulation time: /time

spike train: /extracellular/electrode_1/firing/origin_data/neuron_type/neuron_number

neuron_type is type of neurons, for example, pyr, aac, pvbas

neuron_number is number of neuron in simulation, for example, "neuron_1"

lfp: /extracellular/electrode_1/lfp/origin_data/channel_number

channel_number is number of channel, for example, "channel_1", 1 is stratum oriens, 10 is stratum lacunosum-moleculare

somatic potential: /intracellular/origin_data/neuron_number

neuron_number is the same as for spike train, type of neuron has been noted in attributes.