Interconnectivity

class cell_connection

Describes a connection between two cells: a pre-synaptic source and a post-synaptic target. The source is typically a threshold detector on a cell or a spike source. The target is a synapse on the post-synaptic cell.

The target does not include the gid of a cell, this is because a cell_connection is bound to the target cell, which means that the gid is implicitly known.

cell_global_label_type source

Source end point, represented by a cell_global_label_type which packages a cell gid, label of a group of sources on the cell, and source selection policy.

cell_local_label_type target

Destination end point on the cell, represented by a cell_local_label_type which packages a label of a group of targets on the cell and a selection policy. The target cell’s gid is implicitly known.

float weight

The weight delivered to the target synapse. The weight is dimensionless, and its interpretation is specific to the synapse type of the target. For example, the expsyn synapse interprets it as a conductance with units μS (micro-Siemens).

units::quantity delay

Delay of the connection (milliseconds), must be positive and finite.

class ext_cell_connection

Describes a connection between two cells: a pre-synaptic source and a post-synaptic target. The source is typically a threshold detector on a cell or a spike source. The target is a synapse on the post-synaptic cell.

The target does not include the gid of a cell, this is because a ext_cell_connection is bound to the target cell, which means that the gid is implicitly known.

cell_remote_label_type source

Source end point, comprising

• global id, uniquely identifying the source object inside the remote simulation,

• local id, uniquely identifying the location of the source on the source object,

• selection policy, how to map sources (gid, lid) to the targets in Arbor.

Note

The tuple (gid, lid) must be unique for each source. This addressing scheme follows Arbor’s model of multiple sources (threshold detectors) per cell. If the remote simulation does not provide multiple items per source object, lid can be ignored and set to zero.

cell_local_label_type target

Destination end point on the cell, represented by a cell_local_label_type which packages a label of a group of targets on the cell and a selection policy. The target cell’s gid is implicitly known.

float weight

The weight delivered to the target synapse. The weight is dimensionless, and its interpretation is specific to the synapse type of the target. For example, the expsyn synapse interprets it as a conductance with units μS (micro-Siemens).

float delay

Delay of the connection (milliseconds).

class gap_junction_connection

Describes a gap junction connection between two gap junction sites. The local site does not include the gid of a cell, this is because a gap_junction_connection is bound to the local cell, which means that the gid is implicitly known.

Note

A bidirectional gap-junction connection between two cells c0 and c1 requires two gap_junction_connection objects to be constructed: one where c0 is the local site, and c1 is the peer site; and another where c1 is the local site, and c0 is the peer site.

cell_global_label_type peer

Peer gap junction site, represented by a cell_local_label_type which packages a cell gid, a label of a group of gap junction sites on the cell, and a site selection policy.

cell_local_label_type local

Local gap junction site on the cell, represented by a cell_local_label_type which packages a label of a group of gap junction sites on the cell and a selection policy. The gid of the local site’s cell is implicitly known.

float weight

unit-less gap junction connection weight.

class network_site_info

A network connection site on a cell. Used for generated connections through the high-level network description.

cell_gid_type gid

The cell index.

cell_kind kind

The cell kind.

cell_tag_type label

The associated label.

mlocation location

The local location on the cell.

mpoint global_location

The global location in cartesian coordinates.

class network_connection_info

A network connection between cells. Used for generated connections through the high-level network description.

network_site_info source

The source connection site.

network_site_info target

The target connection site.

class network_value

A network value, describing the calculation for each connection.

network_value scalar(double value)

A fixed scalar value.

network_value named(std::string name)

A named network value in the network label dictionary.

network_value distance()

The value representing the distance between source and target.

network_value uniform_distribution(unsigned seed, const std::array<double, 2> &range)

A uniform random distribution within [range_0, range_1)

network_value normal_distribution(unsigned seed, double mean, double std_deviation)

A normal random distribution with a given mean and standard deviation.

network_value truncated_normal_distribution(unsigned seed, double mean, double std_deviation, const std::array<double, 2> &range)

A truncated normal random distribution with a given mean and standard deviation. Sampled through an accept-reject method to only return values in [range_0, range_1)

network_value custom(custom_func_type func)

Custom value using the provided function “func”. Repeated calls with the same arguments to “func” must yield the same result.

network_value add(network_value left, network_value right)

Summation of two values.

network_value sub(network_value left, network_value right)

Subtraction of two values.

network_value mul(network_value left, network_value right)

Multiplication of two values.

network_value div(network_value left, network_value right)

Division of two values.

network_value min(network_value left, network_value right)

Minimum of two values.

network_value max(network_value left, network_value right)

Maximum of two values.

network_value exp(network_value v)

Exponential of a given value.

network_value log(network_value v)

Logarithm of a given value.

if_else(network_selection cond, network_value true_value, network_value false_value)

if contained in selection, the true_value is used and the false_value otherwise.

class network_selection

A network selection, describing a subset of all possible connections.

network_selection all()

Select all

network_selection none();

Select none

network_selection named(std::string name);

Named selection in the network label dictionary

network_selection inter_cell();

Only select connections between different cells

network_selection source_cell_kind(cell_kind kind);

Select connections with the given source cell kind

network_selection target_cell_kind(cell_kind kind);

Select connections with the given target cell kind

network_selection source_label(std::vector<cell_tag_type> labels);

Select connections with the given source label

network_selection target_label(std::vector<cell_tag_type> labels);

Select connections with the given target label

network_selection source_cell(std::vector<cell_gid_type> gids);

Select connections with source cells matching the indices in the list

network_selection source_cell(gid_range range);

Select connections with source cells matching the indices in the range

network_selection target_cell(std::vector<cell_gid_type> gids);

Select connections with target cells matching the indices in the list

network_selection target_cell(gid_range range);

Select connections with target cells matching the indices in the range

network_selection chain(std::vector<cell_gid_type> gids);

Select connections that form a chain, such that source cell “i” is connected to the target cell “i+1”

network_selection chain(gid_range range);

Select connections that form a chain, such that source cell “i” is connected to the target cell “i+1”

network_selection chain_reverse(gid_range range);

Select connections that form a reversed chain, such that source cell “i+1” is connected to the target cell “i”

network_selection intersect(network_selection left, network_selection right);

Select connections, that are selected by both “left” and “right”

network_selection join(network_selection left, network_selection right);

Select connections, that are selected by either or both “left” and “right”

network_selection difference(network_selection left, network_selection right);

Select connections, that are selected by “left”, unless selected by “right”

network_selection symmetric_difference(network_selection left, network_selection right);

Select connections, that are selected by “left” or “right”, but not both

network_selection complement(network_selection s);

Invert the selection

network_selection random(unsigned seed, network_value p);

Random selection using the Bernoulli random distribution with probability “p” between 0.0 and 1.0

network_selection custom(custom_func_type func);

Custom selection using the provided function “func”. Repeated calls with the same arguments to “func” must yield the same result. For gap junction selection, “func” must be symmetric (func(a,b) = func(b,a)).

network_selection distance_lt(double d);

Only select within a given distance. This may enable more efficient sampling through an internal spatial data structure.

network_selection distance_gt(double d);

Only select if the distance is greater than a given distance. This may enable more efficient sampling through an internal spatial data structure.

class network_label_dict

Dictionary storing named network values and selections.

network_label_dict &set(const std::string &name, network_selection s)

Store a network selection under the given name

network_label_dict &set(const std::string &name, network_value v)

Store a network value under the given name

std::optional<network_selection> selection(const std::string &name) const

Returns the stored network selection of the given name if it exists. None otherwise.

std::optional<network_value> value(const std::string &name) const

Returns the stored network value of the given name if it exists. None otherwise.

const ns_map &selections() const

All stored network selections

const nv_map &values() const

All stored network values

class network_description

A complete network description required for processing.

network_selection selection

Selection of connections.

network_value weight

Weight of generated connections.

network_value delay

Delay of generated connections.

network_label_dict dict

Label dictionary for named selections and values.

generate_network_connections(recipe, context, decomp)

Generate network connections from the network description in the recipe. Only generates connections with local gids in the domain composition as the target. Does not include connections from the connections_on recipe function.

generate_network_connections(recipe) :no-index:

Generate network connections from the network description in the recipe. Returns all generated connections on every process. Does not include connections from the “connections_on” recipe function.