Central Pattern Generators
PROJECT HALTED
A Central Pattern Generator is a network of neurons responsible of generate periodic structures of neural activity in whatever of its different forms.
The simplest form is called half-center oscillator [1], two neurons coupled with inhibitory synapses and an constant input current, something like the next figure:
After asking Gonzalo G. de Polavieja about how to build a biologically plausible CPG I have built from scratch my own using the model developed by Izhikevich which uses only ~6 floating point operations per simulation step.
Now I use two neurons, exhibiting rebound spiking, reciprocally inhibited with a initial hyperpolarizing current of -15 nA for 6 ms, which can be seen in green on the top of the first plot.
Now the simulation looks quite much real. The states of the CPG are as follow:
- Neuron A receives an external stimuli and start firing after the inhibitory input vanishes.
- Neuron A inhibits Neuron B for ~9 ms
- When Neuron A stops inhibiting B this is triggered and fires.
- Neuron B inhibits A
- Go back to state 1