From spikes to cognitive automata in neural assembly computing

It is well known that nervous systems are constructed by spiking neural networks that execute several algorithms concurrently, and that such spiking neurons work together, forming cell assemblies. Finite-state automata (FSA) are abstract constructions that process information and control machines. As FSA execute systematic processing it is possible to say they execute algorithms. Neural Assembly Computing (NAC) is an approach that describes how assemblies of spiking neurons process information and control behavior. We have shown how NAC constructs algorithms by means of FSA, performed on spiking neural networks. In our view, several FSA, each one executing a task and interacting one another, can explain how agents present intelligent behaviors. Generally, FSA are deterministic machines, so they execute a single task, but what happens when FSA change their behaviors as they pass through experiences? We claim this seems to be a kind of cognitive process. In this work, an introduction to NAC is presented, the construction of FSA in spiking neurons is described, and how FSA change their behaviors as they pass through experiences is demonstrated. The NAC approach can explain how intelligent/cognitive processes happen in spiking neural networks from spikes to adaptive behavior.