Location of membrane conductance changes by analysis of the input impedance of neurons. II. Implementation

Mathematical modeling has shown that it should be possible to determine the electrotonic location of membrane conductance changes in single neurons by analysis of the associated changes in the magnitude of the alternating-current (AC) input impedance. The form of the plot of change in the magnitude of the input impedance as a function of frequency (delta [Zn(f)]) should differ for changes in membrane conductance located at different electrotonic distances from the recording/current-injection site. Due to the axial resistance and the membrane capacitance, the higher frequencies are attenuated with distance to a greater degree than are the lower frequencies. Thus delta [Zn(f)] should drop to zero more rapidly with increasing frequency for distal than for proximal conductance changes. For distal changes in conductance, the sign of the change in the magnitude of the input impedance can even reverse in the higher frequency range, so that increases in conductance would produce increases in impedance. This effect may explain the paradoxical increases in impedance at 100 Hz reported for motor neuron inhibitory postsynaptic potentials. Sine-wave impedance measurements were made in single embryonic chick spinal neurons in tissue culture, and gamma-aminobutyric acid (GABA) was iontophoretically applied alternately to the soma and to a neurite at a measured distance from the soma. The impedance changes produced by the GABA-induced conductance changes were consistent with the expectations from the mathematical modeling, but the results suggest that the axial resistance of the neurites must be quite high in some cases. Distortions due to microelectrode capacitance and stray capacitances in the input stage of single-electrode bridge amplifiers can make sine-wave impedance measurements impossible. This difficulty was eliminated by modifications to the capacity compensation circuit of an active bridge amplifier. Noise and distortion of several other types can also introduce serious errors. Methods for minimizing such problems are discussed. In spite of its limitations, this method can be of great practical value, because it can give the electrotonic location of spontaneously occurring membrane conductance changes in single neurons even when unitary synaptic potentials cannot be resolved. These methods are currently being applied to hippocampal pyramidal cells in vivo to locate conductance changes during the electroencephalogram (EEG) theta-rhythm in rats. In such laminated structures, the determination of the anatomical source of a group of active synapses can be aided by location of the resultant membrane conductance changes.