API Reference#
Complete API reference for brainpy.state, organized by model family. The
library exposes two families that share the same state-based substrate:
BrainPy-style — composable, differentiable neurons, synapses, projections, outputs, plasticity, readouts, and input generators.
NEST-compatible — JAX re-implementations that preserve NEST parameter names and semantics, plus the spatial and network-builder layers.
Both families are usable from the single public brainpy.state namespace.
BrainPy-style#
The native modeling layer: compose neurons, synapses, and projections, and train them end to end with surrogate gradients.
Abstract base classes shared by all models: Dynamics, Neuron,
Synapse.
Spiking neuron models (LIF, ALIF, AdEx, HH, Izhikevich, …).
Synaptic dynamics (Expon, DualExpon, Alpha, AMPA, GABAa, BioNMDA).
Connect neural populations (AlignPostProj, DeltaProj,
CurrentProj, gap junctions, AlignPre/AlignPost helpers).
Convert conductances to currents (COBA, CUBA, MgBlock).
Short-term synaptic plasticity (STP, STD).
Readout layers (LeakyRateReadout).
Spike and current generators (PoissonSpike, SpikeTime,
PoissonInput).
NEST-compatible#
JAX re-implementations preserving NEST parameter names, plus the spatial and network-builder layers. Parity with a live NEST install is documented on the validation status page.
NESTNeuron, NESTSynapse, NESTPlasticity, NESTDevice.
IAF, AdEx, GIF, GLIF, HH, Izhikevich, rate, and binary neurons.
Static synapses, gap junctions, and special connections.
STDP, Tsodyks-Markram STP, and voltage-based learning rules.
Generators, recorders, and detectors.
Spatially-structured layers, distance kernels, masks, and visualization.
Declarative network construction, projections, connection rules, and the simulator.