MEArec: a fast and customizable testbench simulatorfor ground-truth extracellular spiking activity

Summary

Collection of simulated extracellular recordings generated with MEArec (https://github.com/alejoe91/MEArec) for the preprint titled: "MEArec: a fast and customizable testbench simulatorfor ground-truth extracellular spiking activity"

The data are used to generate the figures for the paper. In order to reproduce the figures, first clone and install the source code:

git clone https://github.com/alejoe91/MEArec.git
cd MEArec
pip install .

Change directory to the figures folder, create a data folder, then copy and extract the templates.zip and recordings.zip to that folder.
Then you can run the scripts in the figures folder to generate figure 1 to 8 (excluding figure 2). For example, to generate the plots making up figure 3 you would run:

run figure3.py

The template files in the templates folder can also be used to generate new recordings. For example, this command will generate a tetrode recording using the tetrode templates. Assuming you are in the base folder, run:

mearec gen-recordings -t data/templates/tetrode_templates_30.h5 -d 60 -ne 8 -ni 2

This will generate a 1-minute recording with 8 excitatory and 2 inhibitory cells.

For more information, please refer to the manuscript and the documentation (https://mearec.readthedocs.io/en/latest/).

Dataset details

The dataset contains two .zip files: templates.zip and recordings.zip.

The templates.zip include the following template files:

• templates_30_Neuronexus-32_drift.h5 : 30 drifting templates for each cell model (390 templates in total) on a Neuronexus-32 probe. Used in figures 6 and 8.
• templates_30_tetrode.h5 : 30 templates for each cell model (390 in total) on a linear tetrode probe (tetrode-mea-l). Used in figure 1.
• templates_100_Neuronexus-32.h5 : 100 templates for each cell model (1300 in total) on a Neuronexus-32 probe. Used in figures 3, 5 and 8.
• templates_100_Neuropixels-128.h5 : 100 templates for each cell model (1300 in total) on a Neuropixels-128 probe. Used in figure 3.
• templates_100_SqMEA-10-15.h5 : 100 templates for each cell model (1300 in total) on a SqMEA-10-15probe. Used in figure 3.
• templates_300_tetrode_minamp0.h5 :  300 templates for each cell model (3900 in total) on a linear tetrode probe (tetrode-mea-l), with a minimum template amplitude of 0. Used in figure 7 to generate the biological far-neurons noise.

The recordings.zip include the following template files:

• recordings_4cells_Neuronexus-32_60.0_10uV_drift.h5 : drifting recordings with 4 cells on a Neuronexus-32 probe, duration  60 s, noise level 10 uV. Used in figure 6.
• recordings_6cells_tetrode_30.0_10.0uV.h5 : recordings with 6 cells on a linear tetrode probe (tetrode-mea-l), duration 30 s, noise level 10 uV. Used in figures 1 and 4.
• recordings_20cells_Neuronexus-32_30.0_10.0uV.h5 : recordings with 20 cells on a Neuronexus-32 probe, duration 30 s, noise level 10 uV. Used in figure 3.
• recordings_50cells_SqMEA-10-15_30.0_10.0uV.h5 : recordings with 50 cells on a SqMEA-10-15 probe, duration 30 s, noise level 10 uV. Used in figure 3.
• recordings_60cells_Neuropixels-128_30.0_10.0uV.h5 : recordings with 60 cells on a Neuropixels-128 probe, duration 30 s, noise level 10 uV. Used in figure 3.