Examples

Nengo creates these models using the principles of the Neural Engineering Framework. The first set of examples explains these three principles: representation, transformation, and dynamics.

For a summary of these principles, see the following example:

• NEF summary

The following examples give a more detailed presentation:

Principle 1: Representation

• A single neuron
• Two neurons
• Many neurons
• 2-dimensional representation
• Combining
• Addition

Principle 2: Transformation

• Communication channel
• Squaring the input
• Multiplication

Principle 3: Dynamics

• Integrator
• Controlled integrator
• Controlled integrator 2
• A simple harmonic oscillator
• Controlled oscillator
• The Lorenz chaotic attractor

Putting these three principles together allows us to scale these examples up to larger networks that do more complex functions. Below are some of these complex functions, as well as other examples that we hope will be helpful as reference when building your own Nengo models.

Nodes

• Delaying a connection with a node

Processes

• Processes and how to use them

Ensembles

• Tuning curves
• Using the Izhikevich neuron model

Connections

• Inhibitory gating of ensembles
• Improving function approximation by adjusting tuning curves

Learning

• Learning a communication channel
• Learning to square the input
• Learning to compute a product
• Unsupervised learning
• Learning new associations

Networks

• Ensemble array
• Matrix multiplication
• The basal ganglia
• Integrator

Under the hood

• Adding new objects to Nengo
• The NEF algorithm

Reference

• Exceptions
• String representations