Identification of a neuroelectric system involving a single input and a single output

In this work, we identify a neuroelectric system by using a stochastic mode l of Volterra type. The neuroelectric system is called muscle spindle and p lays a critical role in the initiation of movement and in the maintenance o f posture. In order to identify the stochastic model we use spectral analys is techniques of stationary point processes, which are based on Welch's met hod. New asymptotic results for the gain, phase and impulse function of the system based on the Welch's method are obtained. These results are used in examining the behaviour of the muscle spindle under two different experime ntal conditions: (a) when there is no input present and (b) when an input i s present. The presence of the input drastically modifies the behaviour of the muscle spindle. It is shown from the estimates of the gain and phase th at the system behaves as a high-pass filter with the input leading the outp ut by about 30 ms. This result is also verified from the estimate of the im pulse function which indicates that the system does not respond for about 3 0 ms. (C) 2000 Elsevier Science B.V. All rights reserved.