Noisy Dendrites: Models of Dendritic Integration In Vivo

Abstract

While dendritic processing has been well characterized in vitro, there is little experimental data and models available about the integrative properties of dendrites in vivo. Here, we review existing computational models to infer the dendritic processing of neocortical pyramidal neurons in vivo. We start by summarizing experimental measurements of the “high-conductance states” of cortical neurons in vivo. Next, we show models predicting that, in such states, the responsiveness of cortical neurons should be greatly enhanced, in particular due to the presence of high-amplitude fluctuations (“synaptic noise”). We infer that in dendrites this effect should be particularly strong, leading to the spontaneous activation of dendritic spikes. The presence of noise in dendrites also enhances spike propagation. We show that opposite distance dependencies of spike initiation and propagation result in roughly location-independent synaptic efficacies. In addition, in high-conductance states, dendrites display sharper temporal processing capabilities. Thus, we conclude that noisy active dendrites behave more “democratically,” and that dendrites should have enhanced processing capabilities in vivo.