The Journal of Neurophysiology Vol. 87 No. 1 January 2002, pp. 149-156
Copyright ©2002 by the American Physiological Society

Jittery Trains Induced by Synaptic-Like Currents in Cerebellar Inhibitory Interneurons

Department of Neurobiology, Institute of Life Sciences and the Interdisciplinary Center for Neuronal Computation, Hebrew University, Jerusalem 91904, Israel

Mann-Metzer, Puah and Yosef Yarom. Jittery Trains Induced by Synaptic-Like Currents in Cerebellar Inhibitory Interneurons. J. Neurophysiol. 87: 149-156, 2002. Cerebellar inhibitory interneurons respond to parallel fiber input with a characteristic train of action potentials. Here we show that the characteristics of these trains reflect the intrinsic properties of the interneurons. In in vitro cerebellar slices, the response of these neurons to synaptic-like current resembles their in vivo response to parallel fiber inputa train of action potentials characterized by a gradual increase in interspike interval and spike amplitude. A large variability in spike timing, or jitter, was observed, the last action potential emerging from a slow depolarizing wave that lasted beyond the synaptic current and was prevented by either TTX or membrane hyperpolarization. While response duration was weakly dependent on current intensity, the variability of the overall duration was closely related to the variability of the timing of the last action potential. Blocking the Ca²⁺ currents or partial blockade of the delayed rectifier (TEA 2 mM) decreased the excitability, leading to a decrease in the duration and variability of the response and increasing its dependence on stimulus intensity. Increased duration and variability was observed in the presence of Cs⁺ ions (5 mM) that blocked an h-like current. We conclude that a persistent Na⁺ current governs the duration of the response, whereas the synaptic current and the spiking mechanism shape its pattern. The large variability between trials is due to the stochastic nature of the persistent Na⁺ current. Thus unless precise timing is achieved by a network of interconnected neurons, these results vote against temporal coding as a player in the cerebellar computational processing.