A CTRNN is a neural network(NN) that allows recurrent(R) connections and considers continuous time(CT) a factor. It is therefore within the class of dynamic neural networks.
As such, the neurons in a CTRNN may change state over time, with or without any interaction from the external environment. This allows CTRNNs to for example simulate short-time memory or provide fault-tolerant sensor readings.
CTRNNs are by many considered to strike the proper balance between real-world cell-complexity and model simplicity: Complex enough to account for the real world, and simple enough to be effectively simulated and understood. The simple encoding of parameters also make them an optimal substrate for evolutionary algorithms (e.g. genetic algorithms), and especially well-suited for artificial life.
CL-CTRNN makes this functionality available through a CLOS API.
AI—see also: cl-fann, MGL-GPR, GECO