Abstract:Neurons in central nervous systems receive multiple synaptic inputs and transform them into a largely standardized output to their target cells-the action potential. A simplified model posits that synaptic signals are integrated by linear summation and passive propagation towards the axon initial segment, where the threshold for spike generation is either crossed or not. However, multiple lines of research during past decades have shown that signal integration in individual neurons is much more complex, with important functional consequences at the cellular, network, and behavioral-cognitive level. The interplay between concomitant excitatory and inhibitory postsynaptic potentials depends strongly on the relative timing and localization of the respective synapses. In addition, dendrites contain multiple voltage-dependent conductances, which allow scaling of postsynaptic potentials, non-linear input processing, and compartmentalization of signals. Together, these features enable a rich variety of single-neuron computations, including non-linear operations and synaptic plasticity. Hence, we have to revise over-simplified messages from textbooks and use simplified computational models like integrate-and-fire neurons with some caution. This concept article summarizes the most important mechanisms of dendritic integration and highlights some recent developments in the field.

A dendrite is a dendrite is a dendrite? Dendritic signal integration beyond the "antenna" model