Set network structure
set.network.structure.RdThis function sets the structure of a network object, defining its layers, columns, neuron types, and local connectivity parameters. It also generates local nodes based on the specified structure.
Usage
set.network.structure(
network,
neuron_types = c("principal"),
neuron_type_valences = c(1),
layer_names = c("layer"),
n_layers = 1,
n_columns = 1,
layer_height = 1,
column_width = 1,
layer_separation_factor = 3,
column_separation_factor = 3.5,
neurons_per_node = 10,
recurrence_factors = 0.5,
pruning_threshold_factor = 0.1
)Arguments
- network
Network object to configure.
- neuron_types
Character vector giving types of neurons in the network, e.g. c("principal", "interneuron").
- neuron_type_valences
Numeric vector giving valences of each neuron type, e.g. c(1, -1) for excitatory and inhibitory neurons.
- layer_names
Character vector giving names of layers in the network, e.g. c("L2/3", "L4", "L5", "L6").
- n_layers
Integer giving number of layers in the network.
- n_columns
Integer giving number of columns in the network.
- layer_height
Numeric giving height of each layer (in units specified at network creation, default unit is microns, default value is 1.0).
- column_width
Numeric giving width of each column (in units specified at network creation, default unit is microns, default value is 1.0).
- layer_separation_factor
Numeric giving mean distance between layers as a fraction of layer height (default: 3.0).
- column_separation_factor
Numeric giving mean distance between columns as a fraction of column width (default: 3.5).
- neurons_per_node
Matrix giving number of neurons of each type per node in each layer; dimensions must match n_layers (rows) and length of neuron_types (columns).
- recurrence_factors
List of matrices giving local recurrence factors for each layer; each matrix must have dimensions matching length of neuron_types (rows and columns).
- pruning_threshold_factor
Numeric giving factor for pruning weak connections within nodes; connections with strength below this factor times the maximum connection strength in the node will be pruned (default: 0.1).