Skip to content

Simulations

NeuroSeq runs Brian2 spiking network simulations. The model, parameters, input arrays and recorded monitors are saved together in the .nseq project.

Action What it does
Simulate → Run Simulation (Ctrl+R) Asks for a duration and continues from the last simulated time. Runs the external simulator instead if one is connected.
Simulate → Store Simulation Checkpoint (Ctrl+Shift+S) Saves the network state (variables, weights, clock) under a name; pick an existing name to overwrite it.
Simulate → Restore Simulation Checkpoint (Ctrl+Shift+R) Rolls the network back to a chosen checkpoint and truncates the monitors. Checkpoints are saved with the project.
File → Edit Model (Ctrl+E) Parameter tree for equations, parameters, synapses and monitors. Applies on the next run.

Design the input first with the static input editor or a dynamic input routine.

Choosing a code generation backend

Settings → Brian2 Simulation Backend:

  • cython: compiled, fastest; needs a C toolchain (see installation for Windows).
  • numpy: pure Python, always works, slower.

The backend cannot be changed while a simulation runs. For large synapse matrices see Settings → Synapse Matrix RAM Limit (GUI overview).