In neuronal systems, representation of information is distributed among neurons and across cell assemblies. In addition, neuronal systems are characterized by dynamically changing states on different spatio-temporal scales and primarily related to intrinsic (i.e. spontaneous) activity (and plasticity), thus partly confounding our attempts in decoding from neuronal ensemble recordings (extraction of information)
Computational models need to be strongly inspired and linked to the actual dynamics of neuronal assemblies and be able to “talk” with their natural counterpart. The interaction with the natural neuronal systems implies the ability to deal with multiple-input and multiple-output transformation of spatiotemporal patterns of activity (cell assembly replacement).