Uni- and multi-variate coherence function: Utility and misconceptions

The coherence function (CF) is often used to assess the significance of linear relationships between cardiovascular variables, but the literature on applications to baroreceptor sensitivity and dynamic cerebral autoregulation studies shows statistical limitations in its validation. As a consequence, biases can be introduced due to rejection of useful data and/or study participants. Multivariate analysis can extend the utility of CF by allowing the identification of the main variables contributing to complex regulatory control mechanisms and also as a necessary step in the analysis of nonlinear systems.