Signal attenuation as a function of integrator time constant and signal duration
This project was partially supported by the W. C. Archie Fund for Faculty Excellence and the Z. Smith Reynolds Research Leave Fund at Wake Forest University and was presented at the Society for Psycho-phystological Research Meetings, October 1992, San Diego, CA. I thank W. Keith Berg for his comments on an earlier version of this paper.
Abstract
In some laboratories, cyeblink electromyogram is rectified and smoothed before the analog signal is digitized. The relationship among the integrator time constant, the duration of the response, and the attenuation of the data was investigated by passing a number of signals at durations of 5, 10, 20, 50, and 100 ms into a contour-following integrator (Coulbourn model S76–01), which first rectifies and then smooths (integrates) the signal. Integrator time constants of 10, 42, 70, 144, and 300 ms were used. The output of the integrator was sampled (1,000 Hz) and stored on a Macintosh SE computer. The integrator attenuated the signal, and the amount of attenuation increased as the duration of the signal decreased and as the time constant of the integrator increased. For brief inputs, the integrator does not charge fully, so only the early portion of the charge limb is present, and this early portion is the steepest part of the charge limb. Therefore, the longer the time constant, the less likely that small or brief responses will be detected.
