A multitransmitter neural ensemble is a group of neurons interacting not via isolated synaptic connections, but via the emission of neurotransmitters directly into the shared extracellular space (ECS). There are multiple experimental evidence that non-synaptic interactions play the important role in biological neural circuits. We propose a model of multitransmitter neural ensembles where each neuron is represented as a finite state machine. An algorithm of neural interactions via the shared ECS is proposed. This framework allows one to capture the variety of spiking behavior observed in biological neurons. The model is intended primarily for simulation of simple neural ensembles where each neuron has a unique internal properties and plays the specific role in the ensemble activity. We show how the model can imitate such neural activity classes as tonic spiking, bursting, post-inhibitory rebound etc. To illustrate the key features of the proposed framework, we have modeled two examples of pattern-generating neural ensembles: a half-center oscillator and a feeding network of a pond snail.