Amorphic spiking neural network

Hello, this is a totally amorphic examle of spiking neural network. The way it has been built consists in connecting randomly placed neurons - 80% excitatory and 20% inhibitory - with synapses. The total numbers of neurons is 200 and the connections include conduction delay. The more the distance the more the spike takes time to propagate; the delay of the conduction has been implemented easily with a segmented tube (long as distance) in chunks; the conduction is about 10 to 20 chunks. The inhibitory connections don't change (the orange ones) and the excitatory get greener as they get stronger. The thickness of the excitatory connection shows the derivative of the connection itself; in this example you can see a randomly generated pattern applyed so to change the connections; when the pattern is off the net goes on firing neurons creating a wave cycling effect.

Видео Amorphic spiking neural network канала MultiNeurons