VOLTAGE-CLAMP AND SPACE-CLAMP ERRORS ASSOCIATED WITH THE MEASUREMENT OF ELECTROTONICALLY REMOTE SYNAPTIC EVENTS

1. The voltage- and space-clamp errors associated with the use of a so matic electrode to measure current from dendritic synapses are evaluat ed using both equivalent-cylinder and morphologically realistic models of neuronal dendritic trees. 2. As a first step toward understanding the properties of synaptic current distortion under voltage-clamp cond itions, the attenuation of step and sinusoidal voltage changes are eva luated in equivalent cylinder models. Demonstration of the frequency-d ependent attenuation of voltage in the cable is then used as a framewo rk for understanding the distortion of synaptic currents generated at sites remote from the somatic recording electrode and measured in the voltage-clamp recording configuration. 3. Increases in specific membra ne resistivity (R(m)) are shown to reduce steady-state voltage attenua tion, while producing only minimal reduction in attenuation of transie nt voltage changes. Experimental manipulations that increase R(m) ther efore improve the accuracy of estimates of reversal potential for elec trotonically remote synapses, but do not significantly reduce the atte nuation of peak current. In addition, increases in R(m) have the effec t of slowing the kinetics of poorly clamped synaptic currents. 4. The effects of the magnitude of the synaptic conductance and its kinetics on the measured synaptic currents are also examined and discussed. The error in estimating parameters from measured synaptic currents is gre atest for synapses with fast kinetics and large conductances. 5. A mor phologically realistic model of a CA3 pyramidal neuron is used to demo nstrate the generality of the conclusions derived from equivalent cyli nder models. The realistic model is also used to fit synaptic currents generated by stimulation of mossy fiber (MF) and commissural/associat ional (C/A) inputs to CA3 neurons and to estimate the amount of distor tion of these measured currents.6. Anatomic data from the CA3 pyramida l neuron model are used to construct a simplified two-cylinder CA3 mod el. This model is used to estimate the electrotonic distances of MF sy napses (which are located proximal to the soma) and perforant path (PP ) synapses (which are located at the distal ends of the apical dendrit es) and the distortion of synaptic current parameters measured or thes e synapses. 7. Results from the equivalent-cylinder models, the morpho logical CA3 model, and the simplified CA3 model all indicate that the amount of distortion of synaptic currents increases steeply as a funct ion of distance from the soma. MF synapses close to the soma are likel y to be subject only to small space-clamp errors, whereas MF synapses farther from the soma are likely to be substantially attenuated. Synap tic currents from more remote synapses such as C/A and PP inputs are s hown to be enormously attenuated. 8. In conclusion, we show that despi te experimental manipulations to eliminate somatic leak conductances a nd increase R(m), synaptic currents generated in neuronal dendrites an d measured at the soma can still be significantly attenuated and disto rted. Estimates of synaptic conductances and kinetics from voltage-cla mp measurements made at the soma should therefore be paired with estim ates of the errors associated with these measurements. Such estimates will require a knowledge of the location and kinetics for the synapse under study as well as the electrotonic structure of the postsynaptic neuron. The unclamped nature of remote synaptic events also raises the possibility that voltage-gated channels in dendrites may be activated by synaptic inputs, even under voltage-clamp conditions.