Adaptive Estimation Of Event-related Bioelectric Signals: Effect Of Misalignment Errors

Abstract

We present a general adaptive system to estimate deterministic recur- rent transient signals time-locked to a stimulus. This system estimates the components of the deterministic signal using a function base and was applied to evoked potentials and ECG signals. In this work we analyse for this adaptive system the effect of misalignment errors in each signal recurrence. The theoretical analysis shows that the er- rors in the occurrence time estimation produce a low-pass filtering effect wih a cutoff frequency function of the error distribution. A simulation study with real ECG signals agree with theoretical results.