Network Preferences

The Network Preferences dialog can be opened by going to File > Network Preferences in a network window, or by using the keyboard shortcut Cmd/Ctrl-,. These preferences are global settings that apply to all networks in the Simbrain workspace.
Colors
The Colors tab allows customization of all visual color elements in the network display:
- Network background color: Background color of the network canvas
- Node hot color: Color for neurons with high positive activation or excitatory polarity
- Node cool color: Color for neurons with low/negative activation or inhibitory polarity
- Line color: Default color for lines, borders, and neutral elements
- Spiking color: Color used to highlight neurons and connections when spiking
- Excitatory color: Color for excitatory synapses (positive weights)
- Inhibitory color: Color for inhibitory synapses (negative weights)
- Zero color: Color for synapses with zero or near-zero weights
- Weight matrix arrow color: Color for arrows indicating weight matrix connections
- Weight matrix boundary color: Color for weight matrix boundaries and outlines
- Synapse group arrow color: Color for arrows indicating synapse group connections
GUI
The GUI tab controls visual display settings and interface behavior:
- Nudge amount: Distance in pixels to move selected objects when using Shift + arrow keys
- Neuron activation decimal places: Number of decimal places to display in neuron activation text
- Tooltip decimal places: Number of decimal places to display in neuron tooltips
- Min synapse size: Minimum visual diameter for synapse circles in pixels
- Max synapse size: Maximum visual diameter for synapse circles in pixels
- Visibility threshold: Hide individual synapses when synapse groups exceed this number of synapses
- Wand radius: Radius in pixels for the wand tool
- Weight matrix target-source format: If yes, each row of a weight matrix corresponds to an output and each column corresponds to an input. If no (source-target format), each row corresponds to an input and each column corresponds to an output
- Matrix image max width/height: Maximum size of the weight matrix image in pixels in either dimension (width or height)
Model
The Model tab contains parameters that affect network simulation behavior:
- Default network time step: Default time step for network simulation in milliseconds. Neurons that numerically integrate differential equations use this as their integration time step. When the network is in continuous mode, each iteration advances by this value
- Default learning rate: Default learning rate for learning algorithms
Connections
- Default connection strategy: Connection strategy used when pressing 3 in the 1-3 trick. Options include All to All, Sparse, One to One, and other connection strategies
Randomizers
The Randomizers tab configures the probability distributions used when randomizing network elements:
- Weight randomizer: Default randomizer for all free weights and weight matrices, regardless of polarity. Used by:
Randomize selectionmenu item (keyboard shortcut R)Randomize symmetricmenu item for weight matrices- Quick randomization operations throughout the network
This randomizer can produce both positive and negative values, potentially changing the polarity of connections. For weight matrices, consider using the weight matrix histogram dialog for more specialized initialization strategies.
-
Excitatory randomizer: Randomizer for all weights from polarized excitatory neurons. Applying this will not change the polarity of a neuron
-
Inhibitory randomizer: Randomizer for all weights from polarized inhibitory neurons. Applying this will not change the polarity of a neuron
-
Activation randomizer: Randomizer for initial neuron activation values
- Bias randomizer: Randomizer for all neuron biases
Each randomizer can be configured with different probability distributions (Normal, Uniform, Log-Normal, etc.) and their parameters. Click the details button to configure distribution-specific settings.