A new analytical method of studying post-synaptic currents

• Published in: Mathematical biosciences 1999-10, Vol.161 (1), p.15-41
• Main Authors: Poskitt, D.S., Doğançay, K., Chung, Shin-Ho
• Format: Article
• Language: English
• Subjects: Amplitude histogram deconvolution; Biological and medical sciences; Blind deconvolution; Computer Simulation; Dendrites - physiology; Excitatory Postsynaptic Potentials - physiology; Fundamental and applied biological sciences. Psychology; Humans; Isolated neuron and nerve. Neuroglia; Kernel smoothing; Likelihood Functions; Model of synaptic currents; Models, Neurological; Neurons - physiology; Post-synaptic currents; Presynaptic Terminals - physiology; Quantal analysis; Synaptic Transmission - physiology; Vertebrates: nervous system and sense organs
• ISSN: 0025-5564; 1879-3134
• DOI: 10.1016/S0025-5564(99)00038-3

A powerful methodology for analyzing post-synaptic currents recorded from central neurons is presented. An unknown quantity of transmitter molecules released from presynaptic terminals by electrical stimulation of nerve fibers generates a post-synaptic response at the synaptic site. The current induced at the synaptic junction is assumed to rise rapidly and decay slowly with its peak amplitude being proportional to the number of released transmitter molecules. The signal so generated is then distorted by the cable properties of the dendrite, modeled as a time-invariant, linear filter with unknown parameters. The response recorded from the cell body of the neuron following the electrical stimulation is contaminated by zero-mean, white, Gaussian noise. The parameters of the signal are then evaluated from the observation sequence using a quasi-profile likelihood estimation procedure. These parameter values are then employed to deconvolve each measured post-synaptic response to produce an optimal estimate of the transmembrane current flux. From these estimates we derive the amplitude of the synaptic current and the relative amount of transmitter molecules that elicited each response. The underlying amplitude fluctuations in the entire data sequence are investigated using a non-parametric technique based on kernel smoothing procedures. The effectiveness of the new methodology is illustrated in various simulation examples.