A spectral analysis of the integration of artificial synaptic potentials in mammalian central neurons.

In order to simulate the interaction between synaptic input and intrinsic membrane properties in mammalian central neurons a well-defined current was injected into the neurons through a recording electrode. The stimulus was white noise bandpass filtered at 0.5 and 75 Hz and the power spectra of the responses were calculated. Recordings were obtained from neurons of the neocortex, the hippocampus, the thalamus and the cerebellar cortex. The neurons were either located in newly cut slices from adult guinea pig brains or in 3-10-weeks-old slice cultures from brains of newborn rats. In hippocampal and cortical cells the passive membrane properties dominated the shape of the power spectra. In general, when the average membrane potential was made more positive the power of the response increased. When neurons had active subthreshold responses like delayed rectification, sag-and-hump responses or delayed depolarization there was a depression of the response power at frequencies below 10-20 Hz. The depression was voltage dependent in the same way as the current that produced the active subthreshold response. In thalamic cells with a low-threshold Ca2+ spike (lts) the power of the responses grew in the 3-20-Hz range with hyperpolarization. The spectra of the responses of thalamic neurons had multiple peaks indicating multiple frequencies of resonance. Purkinje cells of the cerebellar cortex have prominent plateau potentials. When these cells were stimulated with the white noise at levels where the plateau potentials could be activated the spectra were dominated by a large peak at the lowest frequencies, i.e., below 5 Hz. Few cells in our data base generated spontaneous membrane potential oscillations. When the current stimulus was injected into such neurons the intrinsic rhythm was unaffected by the input and the power spectrum showed a marked peak at the frequency of the intrinsic oscillations. We conclude that bandpass filtered white noise as simulation of synaptic input is valuable for quantification of how passive and active membrane properties affect synaptic integration. The technique can also provide information on the role of transmitters and modulators in the CNS.