This extension defines two NWB neuorodata_types, CompartmentSeries and Compartments. CompartmentSeries stores continuous data (e.g. membrane potential, calcium concentration) from many compartments of many cells, and scales to hundreds of thousands of compartments. Compartments stores the meta-data associated with those compartments, and is stored in SimulationMetaData. Collaboration with Stanford University and the Allen Institute. Funded by the Ripple U19.

pip install ndx-simulation-output

from pynwb import NWBHDF5IO, NWBFile
from datetime import datetime
from ndx_simulation_output import CompartmentSeries, Compartments, SimulationMetaData
import numpy as np


compartments = Compartments()
compartments.add_row(number=[0, 1, 2, 3, 4], position=[0.1, 0.2, 0.3, 0.4, 0.5])
compartments.add_row(number=[0], position=[np.nan])

nwbfile = NWBFile('description', 'id', datetime.now().astimezone())

nwbfile.add_lab_meta_data(SimulationMetaData(compartments=compartments))
cs = CompartmentSeries('membrane_potential', np.random.randn(10, 6),
                       compartments=compartments, unit='V', rate=100.)
nwbfile.add_acquisition(cs)

with NWBHDF5IO('test_compartment_series.nwb', 'w') as io:
    io.write(nwbfile)

conversion from SONTATA:

from ndx_simulation_output.io.from_sonata import sonata2nwb
sonata2nwb('path_to_data_dir', 'nwb_path')

from ndx_simulation_output.io.to_sonata import nwb2sonata
nwb2sonata('nwb_path', 'data_dir2')

command line:

git clone https://github.com/bendichter/ndx-simulation-output.git

in matlab:

generateExtension('/path/to/ndx-simulation-output/spec/ndx-simulation-output.namespace.yaml');

nwb = nwbfile()

[number, number_index] = util.create_indexed_column( ...
    {[0, 1, 2, 3, 4], 0}, '/acquisition/compartments/number');

[position, position_index] = util.create_indexed_column( ...
    {[0.1, 0.2, 0.3, 0.4, 0.5], 0}, '/acquisition/compartments/position');

compartments = types.ndx_simulation_output.Compartments( ...
    'colnames', {'number', 'position'}, ...
    'description', 'membrane potential from various compartments', ...
    'id', types.core.ElementIdentifiers('data', int64(0:5)));

compartments.position = position;
compartments.position_index = position_index;
compartments.number = number;
compartments.number_index = number_index;

membrane_potential = types.ndx_simulation_output.CompartmentSeries( ...
    'data', randn(10,6), ...
    'compartments', types.untyped.SoftLink('/acquisition/compartments'), ...
    'data_unit', 'V', ...
    'starting_time_rate', 100., ...
    'starting_time', 0.0);
    
simulation = types.ndx_simulation_output.SimulationMetaData('compartments', compartments);
    
nwb.general.set('simulation', simulation);

nwb.acquisition.set('membrane_potential', membrane_potential);

Ben Dichter*, Kael Dai*, Aaron Milstein, Yazan Billeh, Andrew Tritt, Jean-Christophe Fillion-Robin, Anton Akhipov, Oliver Ruebel, Nicholas Cain, Kristofer Bouchard, Ivan Soltesz. NWB extension for storing results of large-scale neural network simulations. NeuroInformatics. Montreal, Canada (2018). video.

Dai K, Hernando J, Billeh YN, Gratiy SL, Planas J, et al. (2020) The SONATA data format for efficient description of large-scale network models. PLOS Computational Biology 16(2): e1007696. https://doi.org/10.1371/journal.pcbi.1007696

Developed in collaboration between the Soltesz lab and the Allen Institute during NWB Hackathon #4 by Ben Dichter*, Kael Dai*, Aaron Milstein, Yazan Billeh, Andrew Tritt, Jean-Christophe Fillion-Robin, Anton Akhipov, Oliver Ruebel, Nicholas Cain, Kristofer Bouchard, and Ivan Soltesz