Neuroscience recognizes the problem of the integration of components into a functional whole.
Since 1964, the work of W.J. Freeman has promoted the use of non-linear models in the study of neural interactions in the neocortex, more specifically in perceptual processing involving the olfactory bulb. From his electroencephalogram (EEG) model, Freeman observed that the wavelike process revealed linear and near-equilibrium dynamics. More recently, many scientists involved in neuro-biological research have noticed that, despite extremely non-linear behavior at the microscopic scale, synchronicity emerges, involving many sub-components, in the spatio-temporal context of movement. Synchronicity dissipates as soon as the action is performed. These results hint at a system on the edge of chaos that can switch between states. This would indicate that components can be involved in various tasks based on the type of coupling and most areas are not strictly dedicated to one task.
This type of research has lead to the use of computational models that are reminiscent of the ones developed in statistical mechanics, namely the Ising model, to explore the impact of phase transition in artificial neural networks (ANN).