These cells are the primary output neurons of the cerebellar cortex and cerebellum.

The present system comprises 21 input neurons, connected by optical arrays to a single equivalent output neuron.

These neurons are also the main projection or output neurons of the nucleus accumbens.

The output neurons are arranged according to the given topology.

The weights of output neurons are the only part of the network that can change and be trained.

The weights of output neurons can be learned so that the network can (re)produce specific temporal patterns.

Each of the input neurons connects to every output neuron.

This can be justified by considering the different nature of the non-linear hidden neurons versus the linear output neuron.

Electrical signals tear down a random path from one neuron to the next, and the output neurons say, "red square."

"Getting warmer," the teacher says, so the network adjusts again to a path that's somewhat different and the output neurons say, "purple triangle."