Nonlinear systems analysis of computer models of repetitive firing

✍ Scribed by C. L. Baker

• Publisher: Springer-Verlag
• Year: 1978
• Tongue: English
• Weight: 769 KB
• Volume: 29
• Category: Article
• ISSN: 0340-1200
• DOI: 10.1007/bf00364014

✦ Synopsis

Simulated white noise analysis experiments on a simple integrate-and-fire neuron model with adaptation yields Wiener kernels comparable to those found for a crayfish stretch receptor neuron, for low modulation depths (linear range). At high modulation depths (nonlinear range), this model corresponds well to the neuron only if the "membrane potential" variable is constrained to positive values. An alternative kind of neural model considered is one in which spike initiation processes are ignored, and instead a timecontinuous spike frequency variable is used. Such an analytic differential equation model can be represented by a half-wave rectifier with low-pass feedback; simulated white noise analysis of this model shows good correspondence with the stretch receptor, except at the higher frequencies approaching the cell's carrier frequency. The analytic system model is amenable to mathematical analysis using linear and nonlinear systems theory, resulting in equations which relate features of Wiener kernels (peaks, undershoots, time constants, etc.) to previously described features of neurons (threshold, pacemaker sensitivity, adaptation or post-inhibitory rebound).