This repository contains the code used in the paper "Computational modeling reveals biological mechanisms underlying the whisker-flick EEG". We simulate EEG signals from two versions ("original" and "Schneider-Mizell", respectively) of the Blue Brain Project's models of the rat non-barrel primary somatosensory cortex (nbS1), under a variety of conditions. It relies on the BlueRecording workflow described in this paper and this repo. Briefly, for each of the neural circuits, we generate a "weights file" that describes how the transmembrane current in each neural compartment contributes to the EEG. We then simulate the EEG under a variety of conditions.
We assume that you are running this code on a Linux system with slurm and the spack package manager. This software has not been tested on any other system. Install Neurodamus and BlueRecording according to the instructions in the BlueRecording repository. This repo assumes that your system meets the requirements described there.
Download the model data for the original circuit from this Zenodo repository and extract it into the original_circuit_config folder. Ensure that you use the .asc morphologies.
Copy the model data into the rewired_circuit_config folder. Then replace the edges.h5 file with the version from this Zenodo repository.
To generate the EEG weights file for the original circuit, clone the BlueRecording repository into the parent folder of this repository (to ensure that relative paths work as expected), and follow the instructions here, up to step 4. Alternatively, unzip the file originalCoeffFile.tar.gz from this Zenodo repository into the folder BlueRecording/examples/whiskerFlick/electrodeFile after cloning the repo as described above.
To generate the EEG weights file for the SM circuit, follow the instructions here. Alternatively, unzip the file rewiredCoeffFile.tar.gz from this Zenodo repository into the folder rewired/compartment/6ea6a7b1-3b3b-42c7-868d-8277b09b0597/0/.
The simulations in this repo are organized into "simulation campaigns". Each simulation campaign has 10 subfolders, each corresponding to a simulation with a different random seed.
The simulation campaigns in the folder original refer to the unmodified nbS1 circuit. The simulation campaigns in the folder rewired refer to the nbS1 circuit rewired to incorporate information from the MiCrONS dataset. We refer to this circuit as the Scneider-Mizell (SM) circuit.
First, we run the simulations in the campaigns rewired/97d6aa07-db02-48c6-91c2-b3023ce5bdd0 and original/fixed/a9f782a3-1f22-4384-a122-430bc6b2323c. These are the simulations with full cortico-cortical connectivity, and with thalamic input replayed into the circuit.
To run the simulations, navigate to each simulation folder and launch the launch.sh script.
Once the simulations are complete, postprocess data by launching the Geteeg.sh script in the campaign folder.
To investigate the impact of changing the position of the reference electrode on the EEG signal, run the campaign in rewired/testReference.
To calculate the contribution of presynaptic populations to the EEG signal, we create "spike input files" which list the spike times of the cortical cells from the fully-connected circuits described in the previous section, excluding the spikes from the populations of interest.
To generate these spike input files, run the script EditSpikeFile.sh
We then replay these spike files, along with the thalamic spikes, to a circuit in which spikes generated within the simulation do not trigger synaptic transmission. These simulations are located in the folders disconnected for the original circuit, and rewired/disconnected for the SM circuit.
Run these simulations as described above. The EEG contribution for a specific presynaptic population is the difference between the EEG from the fully-connected circuit and the EEG from these simulations. This difference is calculated in the jupyter notebook described below.
We investigate the effects of compressing the spike times of periTC cells on the EEG signal. We do so by generating two spike input files, one in which all of the spikes from the fully-connected circuit are present except for those from PeriTC cells, and another in which only the spikes from periTC cells are present and where they all occur simultaneously. To generate the first file, run EditSpikeFile_Peri.sh. To generate the second, run EditSpikeTiming.sh
Then, run the simulation campaign in rewired/disconnected/compressedTime.
Rather than running the simulations, you can simply save the postprocessed data from this Zenodo repository into the corresponding folders in this repo.
To reproduce the figures in the paper, simply run the juptyer notebooks in the root directory of this repo.
The development of this software was supported by funding to the Blue Brain Project, a research center of the École polytechnique fédérale de Lausanne (EPFL), from the Swiss government's ETH Board of the Swiss Federal Institutes of Technology.