The prevailing view of neural information processing is based on passive properties of the membrane derived from the application of linear cable theory to dendrites, e.g., [114]. Recent studies, however, have suggested nonlinear models that accommodate new experimental evidence that appears to be inconsistent with this classical view. In particular, conventional theories do not satisfactorily explain: (a) significant fluctuations of synaptic efficacy over short periods of time in response to a recent burst of activation, e.g., [54], (b) a variety of active ion-channels capable of affecting the local membrane electric properties, and (c) nonlinear responses localized to specific dendritic sites, pointing to highly specialized mechanisms correlated with specific inputs. On the other hand, nonlinearities inherent in the new models, e.g., [74,133], give rise to a wider repertoire of (computational) capabilities, such as multiplication, fast correlation, etc. Moreover, experimental and theoretical support for conductive properties of protein filaments and the ionic clouds around them is becoming available, which indicates a far more complex picture of the signaling and information processing phenomena in neurons and other cells.