Build
Train
Learn

Quick Start

Choose a preset to instantly load a network architecture

Dataset

Select the data pattern the network will learn to classify
Data Pattern
The 2D pattern shown in the right panel
Noise
10%
Higher noise = harder classification task

Neuron Model

Configure how neurons compute and communicate
Architecture
Each model processes signals differently
Topology
How layers connect to each other
Activation
Non-linear function applied at each neuron

Layers

Define the network depth and width
Architecture
More layers = deeper network = more complex patterns

Simulation

Control the neural dynamics and learning parameters
Spike Rate
45%
How often neurons fire. Above 85% triggers cascade!
Density
60%
Percentage of possible connections that exist
Time Scale
0.50x
Speed of the simulation (0.05x to 2x)
Threshold
0.75
Voltage needed for a neuron to fire a spike
Learning Rate
0.030
How fast weights update during training

Visual Style

Node Size
0.30
Layer Spacing
10
Node Spacing
1.8
Edges
Labels
Glow
Grid
Weight Colors
STDP Learning
Active Color
Neuron glow
Spike Color
Signal pulse
0 neurons
0 edges
0 spk/s
60 fps
NEURAL CASCADE
Membrane Oscilloscope
STDP Learning Rule

Input Data
0
Neurons
0
Synapses
0
Firing
0
Avg mV
Spike Raster
Membrane Potential
Output Activations
Link copied!

Keyboard Shortcuts

Press ? anytime to toggle this panel

Simulation

SpacePlay / Pause simulation
RRandomize network
TToggle training mode
SStep one frame (when paused)

View

2Toggle 2D / 3D view
OToggle oscilloscope
GToggle grid
AToggle auto-rotate

Navigation

LOpen Learn mode
?Show / hide this help
EscClose any overlay

Mouse

HoverInspect neuron details
ClickPin oscilloscope to neuron
DragRotate camera (3D mode)
ScrollZoom in / out