Parameterized Estimation of Long-Range Correlation and Variance Components in Human Serial Interval Production

Abstract

Repetitive movements lead to isochronous serial interval production which exhibit inherent variability. The Wing-Kristofferson model offers a decomposition of the interresponse intervals in tapping tasks based on a cognitive component and on a motor component. We suggest a new theoretical and fully parametric approach to this model in which the cognitive component is modeled as a long-memory process and the motor component is treated as a white noise process, mutually independent. Under these assumptions, we obtained the autocorrelation function and the spectral density function. Furthermore, we propose an estimator based on the maximization of the frequency-domain representation of the likelihood function. Finally, we conducted a simulation study to assess the properties of this estimator and performed an experimental study involving tapping tasks with two target frequencies (1.250 Hz and 0.625 Hz).