Nelson Spruston and Greg Stuart
Dept. Neurobiology & Physiology,
Institute for Neuroscience, Northwestern University
Division of Neuroscience, John Curtin School of Medical Research., Australian National University
The integration of synaptic inputs by neurons in the central nervous system depends critically on the attenuation of voltage in dendrites. This attenuation depends on at least three factors: 1) the morphology of the dendritic tree, 2) the membrane resistivity, , and 3) the intracellular resistivity, . Direct experimental determination of voltage attenuation are technically difficult; most estimates are therefore inferred from models. Such estimates are generally unreliable, however, because of the uncertainty in the appropriate value of . We have taken advantage of the ability to obtain patch-pipette recordings simultaneously from the soma and dendrites of neurons in brain slices to directly measure voltage attenuation and estimate for pyramidal neurons in layer V of the rat neocortex.
Small hyperpolarizing current pulses were injected at the soma and voltage responses were recorded simultaneously at the soma and in the dendrites. Pooled data from many such recordings indicate that steady-state voltage decays approximately linearly along the apical dendrite, with 50% attenuation at 333 from the soma in control conditions. In the presence of 3-5 mM CsCl to the bathing solution to block ``sag'' in voltage responses, attenuation was considerably less than in control, with 50% attenuation at 545 . This sag is presumably mediated by a Cs-sensitive, hyperpolarization-activated conductance (I). The dramatic effect on voltage attenuation produced by blocking this conductance suggests that I is on at the resting potential and contributes substantially to voltage attenuation in the dendrites.
Neurons were routinely filled with biocytin, and three neurons were fully reconstructed and modeled with NEURON. The membrane time constant, determined from somatic and dendritic voltage responses to long current pulses in the presence of CsCl, was used to constrain the passive membrane properties of the model. In each of these three models, an intracellular resistivity of approximately 150 cm best described the observed voltage attenuation in CsCl. Responses of these models to short current pulses were then compared to experimental data from the same cells. In all three cells, the modeled dendritic voltage rose too slowly to its peak, and the somatic and dendritic voltage responses converged sooner than the experimental data. These observations, along with the observed steady-state voltage attenuation, could only be reconciled using models with lower (70-100 cm) and nonuniform (ranging from about 40,000 cm at the soma to less than 10,000 cm in the distal apical dendrites).
A hyperpolarization-activated sag conductance was also incorporated into the passive models. With this conductance present at uniform density, too little steady-state voltage attenuation was observed from the soma to the apical dendrites. The observed attenuation could only be modeled if a large portion of the total I conductance was concentrated in the apical dendrites. These data suggest that a low intracellular resistivity, as well as nonuniform distributions of I and other channels open at the resting potential play important roles in determining voltage attenuation in layer V pyramidal neurons.