SONATA Simulation Configuration file¶

A simulation configuration file is a json formatted text file used for storing simulation parameters such as stimulus to drive activity, reports to collect data during run, and overrides to adapt the behavior according to experimental conditions and properties.

The file is broken up into separate sections detailed below.

version¶

Optional.

Follows the same guidelines as in Circuit Config Version and shares the same version number. The current version is “2.4”.

manifest¶

Optional.

Variables defining paths and being used in the rest of the simulation configuration file. A variable can be defined as:

“.” which is the path to the directory containing the simulation configuration file
an absolute path

“.” is always resolved as the directory containing the simulation configuration file even in the absence of manifest.

The manifest is only valid in its local file, e.g., the circuit manifest variables are not visible here.

network¶

Optional.

A parameter specifying the path of the circuit configuration file for which the simulation should be performed. The default value is “circuit_config.json”.

example:

"network": "${BASE_DIR}/circuit_config.json"

target_simulator¶

Optional.

A parameter specifying which simulator to run. See Circuit Config target_simulator for valid values. If not specified, the value from the network’s circuit_config.json is used.

node_sets_file¶

Optional.

A file defining the list of nodesets applicable to the simulation. (see SONATA Node Sets) The circuit_config nodesets are considered the base set of nodesets; the values in this file are added to the possible nodesets, and overwrite any duplicates from the base set.

node_set¶

Optional.

A parameter specifying the cells from which node set should be instantiated for the simulation. The absence of that property means that all (non virtual) nodes of all populations are loaded.

compartment_sets_file¶

Optional

Path to a compartment_sets.json file defining precise compartment targets. See File: compartment_sets.json for the full specification.

example:

"compartment_sets_file": "${BASE_DIR}/compartment_sets.json"

run¶

Mandatory.

Parameters defining global simulation settings. As NEURON is the engine used for simulation at BBP, find additional details about Simulation Control at https://nrn.readthedocs.io/en/latest/python/simctrl/programmatic.html

Property	Type	Requirement	Description
tstop	float	Mandatory	Simulation runs until biological time (t) reaches tstop. Given in ms.
dt	float	Mandatory	Duration of a single integration timestep. Given in ms.
random_seed	integer	Mandatory	For random sequences, seed is a positive integer added in order to give the user the capacity to change the sequences.
spike_threshold	float	Optional	The spike detection threshold. A spike is detected whenever the voltage in the spike detection location goes over the spike threshold value. Default is -30.0 mV. NEURON specific details.
integration_method	text	Optional	Selects the NEURON/CoreNEURON integration method. This parameter sets the NEURON global variable h.secondorder. The allowed values are “euler” (default) for fully implicit backward euler, “crank_nicolson” for Crank-Nicolson and “crank_nicolson_ion” for Crank-Nicolson with fixed ion currents.
stimulus_seed	integer	Optional	A non-negative integer used for seeding noise stimuli and any other future stochastic stimuli, default is 0.
ionchannel_seed	integer	Optional	A non-negative integer used for seeding stochastic ion channels, default is 0.
minis_seed	integer	Optional	A non-negative integer used for seeding the Poisson processes that drive the minis, default is 0.
synapse_seed	integer	Optional	A non-negative integer used for seeding stochastic synapses, default is 0.

Note

The electrodes_file field has been moved from the run section to individual LFP report blocks (see SONATA reports). Specifying electrodes_file in the run section is no longer valid.

Migration example:

Before (deprecated):

"run": {
    "tstop": 100,
    "dt": 0.025,
    "random_seed": 12345,
    "electrodes_file": "/path/to/electrodes.h5"
},
"reports": {
    "lfp_report": {
        "type": "lfp",
        "cells": "Mosaic",
        "variable_name": "v",
        "dt": 0.1,
        "start_time": 0,
        "end_time": 100
    }
}

After:

"run": {
    "tstop": 100,
    "dt": 0.025,
    "random_seed": 12345
},
"reports": {
    "lfp_report": {
        "type": "lfp",
        "cells": "Mosaic",
        "dt": 0.1,
        "start_time": 0,
        "end_time": 100,
        "electrodes_file": "/path/to/electrodes.h5"
    }
}

example:

"run": {
     "tstop": 1000,
     "dt": 0.025,
     "random_seed": 201506,
     "integration_method" : 2
}

output¶

Optional.

Parameters to override simulation output.

Property	Type	Requirement	Description
output_dir	text	Optional	Location where output files should be written, namely spikes and reports. Relative paths are interpreted relative to location of simulation_config. Default is relative path ‘output’.
log_file	text	Optional	Specify the filename where console output is written. Default is STDOUT. (When using BBP machines, slurm will capture STDOUT file).
spikes_file	text	Optional	File name where will be listed Action Potentials generated during simulation. Default is out.h5.
spikes_sort_order	text	Optional	The sorting of the Action Potentials. Options include “none”, “by_id”, “by_time”. BBP currently only supports “none” and “by_time”. The default value is “by_time”.

example:

"output": {
     "output_dir": "output",
     "spikes_file": "out.h5"
}

conditions¶

Optional.

Parameters defining global experimental conditions.

Property	Type	Requirement	Description
celsius	float	Optional	Temperature of experiment. Default is 34.0.
v_init	float	Optional	Initial membrane voltage in mV. Default is -80.
spike_location	text	Optional	The spike detection location. Can be either ‘soma’ or ‘AIS’ for detecting spikes in either the soma or axon initial segment, respectively. Default is ‘soma’.
extracellular_calcium	float	Optional	Extracellular calcium concentration. When this parameter is provided, apply it to the synapse uHill parameter to scale the U parameter of synapses (py-neurodamus only feature). If not specified, U is set directly as read from edges file.
randomize_gaba_rise_time	boolean	Optional	When true, enable legacy behavior to randomize the GABA_A rise time in the helper functions. Default is false which will use a prescribed value for GABA_A rise time.
mechanisms		Optional	Properties to assign values to GLOBAL variables in synapse or mechanism MOD files. The format is a dictionary with keys being the SUFFIX names of MOD files (unique names of mechanisms) and values being dictionaries of variable names in the MOD files and their values. Read about NMODL2 SUFFIX description here.
modifications		Optional	List of dictionaries with each member describing a modification that mimics experimental manipulations to the circuit. They are executed in the order as being read from the file.

Parameters required for modifications¶

property	Type	Requirement	Description
name	text	Mandatory	Descriptive name for the modification.
node_set	text	Optional	Node set which receives the manipulation. Either `compartment_set` or `node_set` must be present in each of dictionaries in `modifications` list. Both can’t be missing.
type	text	Mandatory	Name of the manipulation. Supported values are `section_list`, `section`, `compartment_set`, `ttx`, and `configure_all_sections`. `ttx` mimics the application of tetrodotoxin, which blocks sodium channels and precludes spiking. `configure_all_sections` is a generic way to modify variables (properties, mechanisms, etc.) per morphology section. `section_list`, `section` and `compartment_set` are specific manipulation types at different levels of the morphology. See below for more details.
section_configure	text	Mandatory*	For `configure_all_sections` manipulation, a snippet of python code to perform one or more assignments involving section attributes, for all sections that have all the referenced attributes. The wildcard `%s` represents each section. Multiple statements are separated by semicolons. E.g., `%s.attr = value; %s.attr2 \*= value2`. For `section_list`, `section` and `compartment_set` manipulations, a snippet of python code to perform one or more assignments involving attributes as follows: `section_list`: entries should be of the form, e.g. `"apical.gbar_NaTg = 0.0; apical.cm = 1"`. This will set the gbar_NaTg to 0 and cm to 1 for all sections in the apical dendrites. All statements must use the same section list name, such as `apical` in the above example. `section`: entries should be of the form, e.g. `"apic[10].gbar_KTst = 0; apic[10].gbar_NaTg = 0"`. This will set the gbar_KTst to 0 and gbar_NaTg to 0 for all segments of apic[10] section. All statements must reference the same specific section, such as `apic[10]` in the above example. `compartment_set`: entries should be of the form, e.g. `"gbar_KTst = 0; gbar_NaTg = 0"`. This will set the gbar_KTst to 0 and gbar_NaTg to 0 for all segments contained in the property `compartment_set`. Note, when the property `type` is set `compartment_set`, you should also specify the property `compartment_set`. In addition, the simulation must declare the `compartment_sets_file` at the top level. See compartment_sets_file. Sections and segments referenced in the `compartment_sets_file` need not be the same section or section list.
compartment_set	text	Optional	The `compartment_set` to use for manipulation from `compartment_sets_file` json file. The `type` must be `compartment_set`. In the example below, “dend_ca_hotspot_name” is the compartment_set name.

Note

For all modification types (section_list, section, compartment_set), the section_configure is applied only to sections/segments that possess all the referenced attributes. A warning is logged if the configuration applies to zero sections or segments. To reference variables in multiple section types or section indices, multiple modifications dictionaries must be added to the file. User should not combine in a single modifications dictionary. e.g. "type": "section_list", one dictionary for apical sections with "section_configure": "apical.gbar_NaTg = 0.0" and another dictionary for basal sections with "section_configure": "basal.gbar_NaTg = 0.0". Multiple modifications of the same type can be defined in section_configure e.g. "section_configure": "apical.gbar_NaTg = 0.0; apical.cm = 1". If compartment_set is defined, node_set must not be specified. The referenced compartment set must be valid, sorted, and free of duplicates (as in reports). The json file specified by compartment_sets_file is described here in File: compartment_sets.json under SONATA reports.

example:

"conditions": {
     "celsius": 34.0,
     "spike_location": "AIS",
     "mechanisms": {
         "ProbAMPANMDA_EMS": {
             "init_depleted": true,
             "minis_single_vesicle": false
         },
         "ProbGABAAB_EMS" : {
             "property_x": 1,
             "property_y": 0.25
         },
         "GluSynapse": {
             "property_z": "string"
         },
         "StochKv3": {
              "vmin": 0.0
          }
     },
     "modifications": [
         {
             "name": "applyTTX",
             "node_set": "single",
             "type": "ttx"
         },
         {
             "name": "no_SK_E2",
             "node_set": "single",
             "type": "configure_all_sections",
             "section_configure": "%s.gSK_E2bar_SK_E2 = 0"
         },
         {
             "name": "apical_block_NaTg",
             "node_set": "single",
             "type": "section_list",
             "section_configure": "apical.gbar_NaTg = 0"
         },
         {
             "name": "apical[10]_KTst_NaTg_block",
             "node_set": "single",
             "type": "section",
             "section_configure": "apic[10].gbar_KTst = 0; apic[10].gbar_NaTg = 0"
         },
         {
             "name": "Ca_hotspot_dend[10]_manipulation",
             "type": "compartment_set",
             "compartment_set": "dend_ca_hotspot_name",
             "section_configure": "gbar_Ca_HVA2 = 1.5; gbar_Ca_LVA = 2"
         }
     ]
}
"compartment_sets_file": "compartment_sets.json"

The compartment_sets.json should be of the form:

{
  "dend_ca_hotspot_name": {
    "population": "S1nonbarrel_neurons",
    "compartment_set": [
      [1, 35, 0.1],
      [1, 35, 0.3],
      [1, 35, 0.7]
    ]
  }
}

Here, each list in the compartment_set is of the form [node_id, section_id, compartment_id]. See File: compartment_sets.json under SONATA reports for more details. For "name": "Ca_hotspot_dend[10]_manipulation" with [1, 35, 0.1], if section_id 35 represents dend[10] of the neuron node_id 1 in population "S1nonbarrel_neurons", the compartment manipulation will set the gbar_Ca_HVA2 to 1.5 and gbar_Ca_LVA to 2 for the segments dend[10](0.1), dend[10](0.3), and dend[10](0.7).

inputs¶

Optional.

Dictionary of dictionaries with each member describing one pattern of stimulus to be injected.

Property	Type	Requirement	Description
module	text	Mandatory	The type of stimulus dictating additional parameters (see addtional tables below). Supported values: “linear”, “relative_linear”, “pulse”, “sinusoidal”, “subthreshold”, “hyperpolarizing”, “synapse_replay”, “seclamp”, “noise”, “shot_noise”, “relative_shot_noise”, “absolute_shot_noise”, “ornstein_uhlenbeck”, “relative_ornstein_uhlenbeck”, “spatially_uniform_e_field”.
input_type	text	Mandatory	The type of the input with the reserved values : “spikes”, “extracellular_stimulation”, “current_clamp”, “voltage_clamp”, “conductance”. Should correspond according to the module (see additional tables below). Currently, not validated by BBP simulation which will use the appropriate input_type regardless of the string passed.
delay	float	Mandatory	Time in ms when input is activated. Not applicable for the “seclamp” module because the SEClamp is always on at time=0, see NEURON SEClamp.
duration	float	Mandatory	Time duration in ms for how long input is activated.
node_set	text	Optional	Node set which is affected by input. Mutually exclusive with `compartment_set`.
compartment_set	string	Optional	Name of a compartment set from `compartment_sets.json`. Cannot be used with `node_set`. Stimulus will be applied only to the specified compartments.

Note

node_set and compartment_set are mutually exclusive parameters and exactly one of them is required. Providing both or neither is invalid. The simulation must declare the compartment_sets_file at the top level. The referenced compartment set must be valid, sorted, and free of duplicates (as in reports).

Below are additional parameters used depending on the module (input_type)

linear (current_clamp)¶

A continuous injection of current.

Property	Type	Requirement	Description
amp_start	float	Mandatory	The amount of current initially injected when the stimulus activates. Given in nA.
amp_end	float	Optional	If given, current is interpolated such that current reaches this value when the stimulus concludes. Otherwise, current stays at amp_start. Given in nA.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

relative_linear (current_clamp)¶

A continues injection of current, regulated according to the current a cell requires to reach threshold.

Property	Type	Requirement	Description
percent_start	float	Mandatory	The percentage of a cell’s threshold current to inject when the stimulus activates.
percent_end	float	Optional	If given, The percentage of a cell’s threshold current is interpolated such that the percentage reaches this value when the stimulus concludes. Otherwise, stays at percent_start.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

pulse (current_clamp)¶

Series of current pulse injections.

Property	Type	Requirement	Description
amp_start	float	Mandatory	The amount of current initially injected when each pulse activates. Given in nA.
width	float	Mandatory	The length of time each pulse lasts. Given in ms.
frequency	float	Mandatory	The frequency of pulse trains. Given in Hz.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

sinusoidal (current_clamp)¶

Series of current pulse injections.

Property	Type	Requirement	Description
amp_start	float	Mandatory	The peak amplitude of the sinusoid. Given in nA.
frequency	float	Mandatory	The frequency of the waveform. Given in Hz.
dt	float	Optional	Timestep of generated signal in ms. Default is 0.025 ms.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

subthreshold (current_clamp)¶

A continuous injections of current, adjusted from the current a cell requires to reach threshold.

Property	Type	Requirement	Description
percent_less	integer	Mandatory	A percentage adjusted from 100 of a cell’s threshold current. E.g. 20 will apply 80% of the threshold current. Using a negative value will give more than 100. E.g. -20 will inject 120% of the threshold current.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

hyperpolarizing (current_clamp)¶

A hyperpolarizing current injection which brings a cell to base membrance voltage used in experiments. Note: No additional parameter are needed when using module “hyperpolarizing”. The holding current applied is defined in the cell model.

Property	Type	Requirement	Description
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

example:

{
  "inputs": {
    "hypamp_mosaic": {
      "module": "hyperpolarizing",
      "input_type": "current_clamp",
      "delay": 250.0,
      "duration": 1000.0,
      "compartment_set": "apical_dendrites"
    }
  }
}

synapse_replay (spikes)¶

Spike events are created from the cells indicated in a file and delivered to their post synaptic targets. The weights of the replay synapses are set at t=0 ms and are not altered by any delayed connection.

Property	Type	Requirement	Description
spike_file	text	Mandatory	Indicates the location of the file with the spike info for injection. Spikes files are the .h5 spikes files.

seclamp (voltage_clamp)¶

Cells are held at indicated membrane voltage by injecting adapting current.

Property	Type	Requirement	Description
voltage	float	Mandatory	Specifies the initial membrane voltage in mV at which the targeted cells should be held at time = 0. In the case of `duration_levels` and `voltage_levels`, this inital membrane voltage will be overridden by `voltage_level[0]`.
duration_levels	list[float]	Optional	Specifies the durations of each step stimulus. The first step begins at time = 0, and each subsequent step begins at the cumulative sum of all previous durations. The sum of `duration_levels` must not exceed the `duration` property. If the sum of `duration_levels` is less than `duration` , the last value in `voltage_levels` will be maintained until the end. It is recommended that durations be multiples of the simulation time step (dt) (i.e., duration = n·dt for some integer n) to ensure proper alignment with NEURON’s fixed-step method where Vector.play updates occur only at discrete time steps (fadvance()).
voltage_levels	list[float]	Optional	Specifies the membrane voltages the targeted cells should be held at in mV for each step stimulus. `voltage_levels[0]` overrides the `voltage` property.
series_resistance	float	Optional	Specifies the series resistance in M \(\Omega\). Default is 0.01 M \(\Omega\).

noise (current_clamp)¶

Continuous injection of current with randomized deflections. Note: one must chose either “mean” or “mean_percent”.

Property	Type	Requirement	Description
mean	float	Mandatory*	The mean value of current to inject. Given in nA.
mean_percent	float	Mandatory*	The mean value of current to inject as a percentage of a cell’s threshold current.
variance	float	Optional	The variance around the mean of current to inject using a normal distribution.
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON IClamp mechanism. The IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a MembraneCurrentSource mechanism, which is identical to IClamp, but produce a membrane current, which is included in the calculation of the extracellular signal.

example:

"inputs": {
     "threshold_exc": {
          "module": "noise",
          "input_type": "current_clamp",
          "mean_percent": 78,
          "variance": 0.1,
          "delay": 500,
          "duration": 3000,
          "node_set": "L5TTPC"
     }
}

shot_noise, absolute_shot_noise and relative_shot_noise (current_clamp or conductance)¶

Generate a Poisson shot noise signal consisting of bi-exponential pulses with gamma distributed amplitudes occurring at exponentially distributed time intervals, resembling random synaptic input. In the Relative and Absolute versions the three parameters (rate, amp_mean, amp_var) are obtained from other three parameters: (amp_cv, mean_percent, sd_percent) for Relative and (amp_cv, mean, sigma) for Absolute, through and analytical result that connects them. In the Relative version the parameters (mean, sigma) are computed relative to a cell’s threshold current (current_clamp) or inverse input resistance (conductance), by scaling these with (mean_percent, sd_percent). The input resistance values must be provided as an additional dataset @dynamics/input_resistance in the nodes file. Note: fields marked Mandatory* depend on which shot_noise version is selected.

Property	Type	Requirement	Description
rise_time	float	Mandatory	The rise time of the bi-exponential shots in ms.
decay_time	float	Mandatory	The decay time of the bi-exponential shots in ms.
rate	float	Mandatory*	For shot_noise, rate of Poisson events in Hz.
amp_mean	float	Mandatory*	For shot_noise, mean of gamma-distributed amplitudes in nA (current_clamp) or uS (conductance).
amp_var	float	Mandatory*	For shot_noise, variance of gamma-distributed amplitudes in nA^2 (current_clamp) or uS^2 (conductance).
mean_percent	float	Mandatory*	For relative_shot_noise, signal mean as percentage of a cell’s threshold current (current_clamp) or inverse input resistance (conductance).
sd_percent	float	Mandatory*	For relative_shot_noise, signal std dev as percentage of a cell’s threshold current (current_clamp) or inverse input resistance (conductance).
mean	float	Mandatory*	For absolute_shot_noise, signal mean in nA (current_clamp) or uS (conductance).
sigma	float	Mandatory*	For absolute_shot_noise, signal std dev in nA (current_clamp) or uS (conductance).
relative_skew	float	Optional	For relative_shot_noise and absolute_shot_noise, signal skewness as a fraction in [0, 1]. This fraction represents a value between the minimum and maximum skewness values compatible with the given signal mean and std dev. Default is 0.5.
reversal	float	Optional	Reversal potential for conductance injection in mV. Default is 0.
dt	float	Optional	Timestep of generated signal in ms. Default is 0.25 ms.
random_seed	integer	Optional	Override the random seed (to introduce correlations between cells).
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON SEClamp mechanism, if a conductance source, or a NEURON IClamp mechanism, if a current source. The SEClamp and IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a ConductanceSource mechanism or a MembraneCurrentSource mechanism, which are identical to SEClamp and IClamp, respectively, but produce a membrane current, which is included in the calculation of the extracellular signal.

ornstein_uhlenbeck and relative_ornstein_uhlenbeck (current_clamp or conductance)¶

Generate an Ornstein-Uhlenbeck process signal injected as a conductance or current. In the Relative version the parameters (mean, sigma) are computed relative to a cell’s inverse input resistance (conductance) or threshold current (current_clamp), by scaling these with (mean_percent, sd_percent). The input resistance values must be provided as an additional dataset @dynamics/input_resistance in the nodes file. Note: fields marked Mandatory* depend on which ornstein_uhlenbeck version is selected.

Property	Type	Requirement	Description
tau	float	Mandatory	Relaxation time constant in ms.
mean_percent	float	Mandatory*	For relative_ornstein_uhlenbeck, signal mean as percentage of a cell’s threshold current (current_clamp) or inverse input resistance (conductance).
sd_percent	float	Mandatory*	For relative_ornstein_uhlenbeck, signal std dev as percentage of a cell’s threshold current (current_clamp) or inverse input resistance (conductance).
mean	float	Mandatory*	For ornstein_uhlenbeck, signal mean in nA (current_clamp) or uS (conductance).
sigma	float	Mandatory*	For ornstein_uhlenbeck, signal std dev in nA (current_clamp) or uS (conductance).
reversal	float	Optional	Reversal potential for conductance injection in mV. Default is 0.
dt	float	Optional	Timestep of generated signal in ms. Default is 0.25 ms.
random_seed	integer	Optional	Override the random seed (to introduce correlations between cells).
represents_physical_electrode	boolean	Optional	Default is False. If True, the signal will be implemented using a NEURON SEClamp mechanism, if a conductance source, or a NEURON IClamp mechanism, if a current source. The SEClamp and IClamp produce an electrode current which is not included in the calculation of extracellular signals, so this option should be used to represent a physical electrode. If the noise signal represents synaptic input, represents_physical_electrode should be set to False, in which case the signal will be implemented using a ConductanceSource mechanism or a MembraneCurrentSource mechanism, which are identical to SEClamp and IClamp, respectively, but produce a membrane current, which is included in the calculation of the extracellular signal.

spatially_uniform_e_field (extracellular_stimulation)¶

Generates a temporally-oscillating extracellular potential field. The potential field is defined as the sum of an arbitrary number of potential fields which vary cosinusoidally in time, and whose gradient (i.e., E field) is constant.

Property	Type	Requirement	Description
fields	list	Mandatory	A list of dicts, where each dict defines one of the fields which are summed to produce the total stimulus. The format for each such dict is given in the table below.
ramp_up_time	float	Optional	Duration during which the amplitude of the signal ramps up linearly from 0, in ms. If not provided, assume no ramp-up time (note that the specified “duration” parameter is not inclusive of this ramp-up time)
ramp_down_time	float	Optional	Duration during which the amplitude of the signal ramps down linearly to 0, in ms. If not provided, assume no ramp-down time (note that the specified “duration” parameter is not inclusive of the ramp-down time)

Property	Type	Requirement	Description
Ex	float	Mandatory	Peak amplitude of the cosinusoid in the x-direction, in V/m. May be negative
Ey	float	Mandatory	Peak amplitude of the cosinusoid in the y-direction, in V/m. May be negative
Ez	float	Mandatory	Peak amplitude of the cosinusoid in the z-direction, in V/m. May be negative
frequency	float	Optional	Frequency of the cosinusoid, in Hz. Must be non-negative. If not provided, assumed to be 0. In this case, a time-invariant field with amplitude [Ex, Ey, Ez] is applied, unless ramp_up_time or ramp_down_time is specified, in which case the field will increase/decrease linearly with time during the ramp periods, and will be constant during the remaider of the stimulation period. Note that the signal will be generated with the same time step as the simulation itself. Note that frequency should therefore be less than the Nyquist frequency of the simulation (i.e., 1/(2*dt))
phase	float	Optional	Phase of the cosinusoid, in radians. If not provided, assumed to be 0.

reports¶

Optional.

Dictionary of dictionaries with each member describing one data collection during the simulation such as compartment voltage.

Property	Type	Requirement	Description
cells	text	Optional	Specify which node_set to report, default is the simulation “node_set”.
sections	text	Optional	Specify which section(s) to report, available labels are dependent on the model setup. To report on all sections, use the keyword “all”. Default is “soma”. At BBP, we currently support “soma”, “axon”, “dend”, “apic”, or “all”.
type	text	Mandatory	Indicates type of data collected. “compartment”, “summation”, “synapse”, or “lfp”. Compartment means that each compartment outputs separately in the report file. Summation will sum up the values from compartments to write a single value to the report (section soma) or sum up the values and leave them in each compartment (other section types). More on summation after the table. Synapse indicates that each synapse afferent to the reported cells will have a separate entry in the report. LFP will report the contribution to the lfp (or eeg) signal from each cell, using the `electrodes_file` specified in this report block. See more after the table.
scaling	text	Optional	For summation type reporting, specify the handling of density values: “none” disables all scaling, “area” (default) converts density to area values. This makes them compatible with values from point processes such as synapses.
compartments	text	Optional	For compartment type reporting, override which compartments of a section are selected to report. Options are “center” or “all”. When using “sections”:”soma”, default is “center”, for other section options, default is “all”.
variable_name	text	Mandatory*	The Simulation variable to access. The variables available are model dependent. For summation type, can sum multiple variables by indicating as a comma separated strings. e.g. “ina, ik”. Mandatory for all report types except `"lfp"`, where it is not allowed (see LFP report for details on LFP computation).
unit	text	Optional	String to output as descriptive test for unit recorded. Not validated for correctness.
dt	float	Mandatory	Interval between reporting steps in milliseconds. If assigned value smaller than simulation dt, will be set equal to simulation dt.
start_time	float	Mandatory	Time to start reporting in milliseconds.
end_time	float	Mandatory	Time to stop reporting in milliseconds.
file_name	text	Optional	Specify report file name to be written in the output_dir. The ‘.h5’ extension will be added if not provided. The default file name is <report_name>.h5 where ‘report_name’ is the key name of the current dictionary.
enabled	boolean	Optional	Allows for supressing a report so that it is not created. Useful for reducing output temporarily. Possible values are true/false. Default is true.
compartment_set	text	Optional	Name of a compartment set from `compartment_sets.json`. Required if `type` is “compartment_set”.
electrodes_file	text	Mandatory*	Path to the HDF5 weights file containing electrode scaling factors for LFP computation. Format described in SONATA Technical description. Mandatory when `type` is `"lfp"`. Path resolution follows the same rules as other file paths in the Simulation_Config.

Note

cells in reports is similar to the node_set in inputs.

In compartment reports, cells and compartment_set are mutually exclusive parameters, and exactly one of them is required. Providing both or neither is invalid.

Examples:

"reports": {
     "soma": {
          "cells": "Mosaic",
          "sections": "soma",
          "type": "compartment",
          "variable_name": "v",
          "unit": "mV",
          "dt": 0.1,
          "start_time" : 0,
          "end_time" : 500,
          "file_name": "soma"
          "enabled" : true
     },
     "compartment": {
          "cells": "Mosaic",
          "sections": "all",
          "type": "compartment",
          "variable_name": "v",
          "unit": "mV",
          "dt": 0.1,
          "start_time" : 0,
          "end_time" : 500,
          "file_name": "voltage"
          "enabled" : true
     },
     "compartment_set": {
          "compartment_set": "Mosaic_A",
          "type": "compartment_set",
          "variable_name": "v",
          "dt": 0.1,
          "start_time": 1000.0,
          "end_time": 1275.0,
          "unit": "mV"
       },
     "axonal_comp_centers": {
          "cells": "Mosaic",
          "sections": "axon",
          "type": "compartment",
          "variable_name": "v",
          "unit": "mV",
          "compartments": "center",
          "dt": 0.1,
          "start_time" : 0,
          "end_time" : 500,
          "file_name": "axon_centers"
          "enabled" : true
     },
     "cell_imembrane": {
         "cells": "Column",
         "sections": "soma",
         "type": "summation",
         "variable_name": "i_membrane, IClamp",
         "unit": "nA",
         "start_time": 0,
         "end_time": 500,
         "enabled": true
     },
     "dend_report_v": {
      "type": "compartment_set",
      "compartment_set": "example_compartment_set",
      "variable_name": "v",
      "unit": "mV",
      "dt": 0.1,
      "start_time": 0.0,
      "end_time": 100.0
     }
},
"compartment_sets_file": "circuit/compartment_sets.json"

See Fine-grained Compartment report for more details on compartment sets.

connection_overrides¶

Optional.

List of dictionaries to adjust the synaptic strength or other properties of edges between two sets of nodes. These are executed in the order they are read from the file. If a set of synapses are affected by multiple connection_overrides because of source and target used, the latter will overwrite any repeated fields set by a former. This is useful when making more general adjustments and then more specific adjustments. Any edges unaffected by any connection_overrides are instantiated as prescribed in the model.

Property	Type	Requirement	Description
name	text	Mandatory	Descriptive name for the override.
source	text	Mandatory	node_set specifying presynaptic nodes.
target	text	Mandatory	node_set specifying postsynaptic nodes.
weight	float	Optional	Multiplier of `conductance` used to adjust synaptic strength. The final NetCon weight in NEURON is computed as `weight * conductance`, where `conductance` is obtained from the `edges.h5` file.
spont_minis	float	Optional	Synapses affected by this connection_override section will spontaneously trigger with the given rate.
synapse_configure	text	Optional	Provide a HOC snippet for synapse objects in `connection_override`, using `%s` for the synapse reference (e.g. `%s.NMDA_ratio = 1 %s.Fac = 2`). Global variables can also be overridden without `%s` (e.g., `tau_d_NMDA_ProbAMPANMDA_EMS = 3`). but these changes do not account for delays, synapse type, or file order (which may depend on the edge file order). Use with caution, as unintended interactions may occur. In fact, it is strongly suggested to use the method described in conditions to modify global synapse variables. However, this method takes priority. A safe approach is to modify global variables, if needed, only once in the entire file. Their values remain constant throughout the simulation, as all changes are applied at the start.
modoverride	text	Optional	Changes the synapse helper files used to instantiate the synapses in this connection. A synapse helper initializes the synapse object and the parameters of the synapse model. By default, AMPANMDAHelper.hoc / GABAABHelper.hoc are used for excitatory / inhibitory synapses. The value of this field determines the prefix of the helper file to use e.g. “GluSynapse” would lead to GluSynapseHelper.hoc being used. That helper will use the additional parameters of the plastic synapse model read from the SONATA edges file using Neurodamus. This is required when using the GluSynapse.mod model and will fail for other models, or if the parameters are not present in the edges file.
synapse_delay_override	float	Optional	Value to override the synaptic delay time originally set in the edge file, and to be given to netcon object. Given in ms.
delay	float	Optional	Adjustments from weight of this connection_override are applied after specified delay has elapsed in ms. Note that only weight modifications are applied so all other fields (spont_minis, synapse_configure, modoverride, synapse_delay_override) are ignored.
neuromodulation_dtc	float	Optional	Only applicable to NeuroModulation projections. It overrides the `neuromod_dtc` values between the selected source and target neurons, representing the decay time constant of the neuromodulator concentration at the target synapse. Given in ms.
neuromodulation_strength	float	Optional	Only applicable to NeuroModulation projections. It overrides the `neuromod_strength` values between the selected source and target neurons, representing the amount of increase of the neuromodulator concentration at the synapse when an incoming neuromodulatory event (i.e., a spike in the virtual pre-synaptic neuron) is transmitted to the target synapse. Given in \(\mu M\).

example:

"connection_overrides": [
     {
          "name": "weaken_excitation"
          "source": "Excitatory",
          "target": "Mosaic,
          "weight": 0.75,
          "spont_minis": 0.04
     },
     {
          "name": "deactivate_short_term_plasticity",
          "source": "Mosaic",
          "target": "Mosaic",
          "synapse_configure": "%s.Fac = 0 %s.Dep = 0"
     }
]

metadata¶

A set of variables storing remarks on the simulation, but are not used for running the simulation.

example:

"metadata": {
     "note": "the first attempt at reproducing xxx experiment",
     "version": "v1",
     "v_int": 10,
     "v_float": 0.5,
     "v_bool": false
}

beta_features¶

This section is reserved for variables that are used for developing a new feature of the simulation. Once the feature goes in production, the variables should be moved to a proper section in the simulation configuration file.

example:

"beta_features": {
     "v_str": "abcd",
     "v_float": 0.5,
     "v_int": 10,
     "v_bool": false
}