1/ƒ noise and chaos in analog neural networks

A model of neural network with sparsely connected analog neurons is presented. In this conceptualization, the trans-membrane conductance controls the dynamical features of the network activity which can display intermittent, chaotic, periodic or constant patterns. An explanation for intermittency is proposed by assuming that the system is self-critical in this particular phase. A connection is then established with experimental observations from EEG of living subjects.