Recipe description¶

This document specifies all parameters required to generate the connectome of a circuit.

Components of the recipe are used by TouchDetector and Functionalizer as noted. An example or a TouchDetector configuration may look like:

recipe:
  bouton_distances:
    min_distance: 0.1
    max_distance: 10.0
    region_gap: 5.0
  structural_spine_lengths:
    - mtype: L7_A
      spine_length: 5.0
    - mtype: L9_DKE
      spine_length: 12.3

Generic pathway selection¶

One can create rules that only apply to certain “pathways” by restricting attributes of source and target node populations. As such, the rules have properties that are prefixed with src_ to match attributes from the source node population, and likewise prefixed with dst_ to match target node population attributes. All attributes are described in Node File, the currently supported ones are:

etype
mtype
region
synapse_class

All supported attributes support the use of simple wild cards. The following rule matches all synapses connecting cells of MType starting with L6_ to ones with MType L1_PYR, from any source region to the target region SPYR:

- src_mtype: "L6_*"
  src_region: "*"
  dst_mtype: "L1_PYR"
  dst_region: "SPYR"

When a component can be stored as a Pandas DataFrame, all used attributes have to be referred to with a _i suffix and stored as numerical values corresponding to the index of the attribute value in the @library field of the node files. Wildcards are no longer allowed, but a numerical value of -1 may be used to match all possible values.

Warning

Storing recipe components in DataFrames ties the recipe strongly to one circuit. Only if the order in the used @library fields matches exactly may the recipe be reused for a different circuit.

Given MType values of L1_PYR, L2_FOO, L6_SPAM, L6_HAM, L6_EGGS, and region values of BLARGH, SPYR, the above rule then turns into:

src_mtype_i	src_region_i	dst_region_i
3	-1	1
4	-1	1
5	-1	1

Generic Components¶

`version`¶

Required.

An integer representing the current version of the recipe. This document describes version 1.

Components specific to TouchDetector¶

Both components are required when running TouchDetector.

`bouton_interval`¶

Required.

Distances used when transforming touch regions into synapse candidates.

Property	Requirement	Description
`min_distance`	Mandatory	The minimum distance at which two synapse candidates are placed from each other in µm.
`max_distance`	Mandatory	The maximum distance at which two synapse candidates are placed from each other in µm.
`region_gap`	Mandatory	The maximum distance between two touch regions at which they are merged in µm.

`structural_spine_lengths`¶

Required.

A list that specified how long the spines for certain MTypes may be. Requires that all MTypes have a spine length assigned. Each item of the list must have the following properties:

Property	Requirement	Description
`mtype`	Mandatory	The MType to apply the spine length to.
`spine_length`	Mandatory	Maxiumum spine length, in µm.

Components specific to Functionalizer¶

Components used by Functionalizer may only be required when the corresponding filter is used. I.e., if only the SynapseProperties filter is used, only the synapse_properties part of the recipe is required.

`bouton_distances`¶

Optional.

Minimum distances for synapses, as measured along the branch length, starting at the soma. Synapses with less than the specified distance will be removed by Functionalizer.

Property	Requirement	Description
`excitatory_synapse_distance`	Optional	The minimum distance from the soma for a synapse of post-synaptic excitatory cells in µm. Defaults to 5µm.
`inhibitory_synapse_distance`	Optional	The minimum distance from the soma for a synapse of post-synaptic inhibitory cells in µm. Defaults to 25µm.

`connection_rules`¶

Required.

A list of rules that are used to determine how synapse distributions are calculated for pathways, and used to reduce the structural connectome to a functional one.

Each rule may have properties corresponding to Generic pathway selection. In addition to the selection attributes, exactly one set of parameters have to be used:

mean_syns_connection, stdev_syns_connection, and active_fraction
bouton_reduction_factor, cv_syns_connection, and active_fraction
bouton_reduction_factor, cv_syns_connection, and mean_syns_connection
bouton_reduction_factor, cv_syns_connection, and probability
bouton_reduction_factor, pMu_A, and p_A

Where the parameters signify:

Parameter	Description
`active_fraction`	The fraction of synapses to be removed in the third pruning step.
`bouton_reduction_factor`	The fraction of synapses to be removed in all three pruning steps.
`cv_syns_connection`	The target value for the coefficient of variation of the distribution of synapses per connection distribution of synapses per connections.
`mean_syns_connection`	The target value for the mean of the distribution of synapses per connections.
`p_A`	The reduction factor.
`pMu_A`	Used as input to the survival rate.
`probability`	The target connection probability. To be deprecated.
`stdev_syns_connection`	The target value for the standard deviation of the distribution of synapses per connection.

`gap_junction_properties`¶

Optional.

A global default setting for the conductance produced by Functionalizer.

Property	Requirement	Description
conductance	Optional	The conductance to be used by all synapses. Defaults to 0.2.

`seed`¶

Optional.

One of the random number seeds to be used when drawing distributions to cut synapses or determine properties.

`synapse_properties`¶

Required.

Settings to generate synaptic properties for appositions. Each apposition is classified by rules and synaptic properties are generated per cell-cell connection following the parameters of the property configuration.

Property	Requirement	Description
`rules`	Mandatory	Rules to classify synapses
`classes`	Mandatory	Synapse classes used to assign properties

`rules`¶

Each rule may have properties corresponding to Generic pathway selection. In addition to the selection attributes, the following parameters may be present:

Property	Requirement	Description
`class`	Mandatory	A name that will be referenced by `classes`. It has to start with either `E` for excitatory connections or `I` for inhibitory connections.
`neural_transmitter_release_delay`	Optional	Defaults to 0.1 ms
`axonal_conduction_velocity`	Optional	Defaults to 300 μm/ms

Please note that contrary to the legacy XML recipe, precedence is handled very strict: later rules always override earlier ones. There is no special treatment for more general rules to not override more specialized ones.

`classes`¶

Here, the class field has to match a class value of the rules. The properties are assigned using the following random number distributions, using a mean m and standard deviation sd:

A Gamma-distribution, with shape parameter equal to m² / sd², and scale parameter equal to sd² / m.
A truncated Normal-distribution, where values are redrawn until they are both positive and within the range of m±sd.
A Poisson-distribution using only m.

The same drawn number is reused for all synapses within the same source to target cell connection.

The following properties are supported, with the mean specified by the property name, and the standard deviation by appending _sd to the property name:

Property	Requirement	Description
`conductance_mu`	Mandatory	The central value for the peak conductance (in nS) for a single synaptic contact, following a Gamma distribution.
`conductance_sd`	Mandatory	Standard deviation of `conductance`.
`depression_time_mu`	Mandatory	Central value for the time constant (in ms) for recovery from depression, following a Gamma distribution.
`depression_time_sd`	Mandatory	Standard deviation of `depression_time`.
`facilitation_time_mu`	Mandatory	Central value for the time constant (in ms) for recovery from facilitation, following a Gamma distribution.
`facilitation_time_sd`	Mandatory	Standard deviation of `f`.
`u_syn_mu`	Mandatory	Central value for the utilization of synaptic efficacy, following a truncated Normal distribution.
`u_syn_sd`	Mandatory	Standard deviation of `u`.
`decay_time_mu`	Mandatory	Central value for the decay time constant (in ms), following a truncated Normal distribution.
`decay_time_sd`	Mandatory	Standard deviation of `dtc`.
`n_rrp_vesicles_mu`	Mandatory	Central value for the number of vesicles in readily releasable pool, following a Poisson distribution.
`conductance_scale_factor`	Optional	The scale factor for the conductance; SRSF: ‘synaptic receptor scaling factor’.
`u_hill_coefficient`	Optional	A coefficient describing the scaling of u to be done by the simulator.

Truncated Normal distributions are limited to the central value ±σ and are re-rolled until positive values has been obtained.

The u_hill_coefficient is used as follows:

\[u_\text{final} = u \cdot y \cdot \frac{ca^4}{u_\text{Hill}^4 + ca^4}\]

where \(ca\) denotes the simulated calcium concentration in millimolar and \(y\) a scalar such that at \(ca = 2.0:\ u_\text{final} = u\). (Markram et al., 2015)

These attributes will be copied for each synapse corresponding to its classification. If they are not specified, no corresponding columns will be created in the output.

`synapse_reposition`¶

Required.

Property	Requirement	Description
`src_mtype`	Mandatory	The MType of the source cell.
`dst_mtype`	Mandatory	The MType of the target cell.
`class`	Mandatory	Has to be `AIS`.

`touch_rules`¶

Required.

Determines which touches are allowed, depending on source and target node population MType, as well the section type on either the source or target side of the touch.

Property	Requirement	Description
`src_mtype`	Mandatory	The MType of the source cell.
`dst_mtype`	Mandatory	The MType of the target cell.
`afferent_section_type`	Optional	The section type of the target cell
`efferent_section_type`	Optional	The section type of the source cell.

The afferent_section_type and efferent_section_type may take the values soma, axon, apical, and basal. A special value dendrite may be used to signify both apical and basal types.

`touch_reduction`¶

Required.

Used to cut touches according to a flat survival rate set by the user. Affects all touches the same way.

Property	Requirement	Description
`survival_rate`	Mandatory	A flat survival probability of touches.

Recipe description¶

Generic pathway selection¶

Generic Components¶

version¶

Components specific to TouchDetector¶

bouton_interval¶

structural_spine_lengths¶

Components specific to Functionalizer¶

bouton_distances¶

connection_rules¶

gap_junction_properties¶

seed¶

synapse_properties¶

rules¶

classes¶

synapse_reposition¶

touch_rules¶

touch_reduction¶

`version`¶

`bouton_interval`¶

`structural_spine_lengths`¶

`bouton_distances`¶

`connection_rules`¶

`gap_junction_properties`¶

`seed`¶

`synapse_properties`¶

`rules`¶

`classes`¶

`synapse_reposition`¶

`touch_rules`¶

`touch_reduction`¶