RT Journal Article
SR Electronic
T1 The cross-frequency mediation mechanism of intracortical information transactions
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 119362
DO 10.1101/119362
A1 RD Pascual-Marqui
A1 P Faber
A1 S Ikeda
A1 R Ishii
A1 T Kinoshita
A1 Y Kitaura
A1 K Kochi
A1 P Milz
A1 K Nishida
A1 M Yoshimura
YR 2017
UL http://biorxiv.org/content/early/2017/03/23/119362.abstract
AB In a seminal paper by von Stein and Sarnthein (2000), it was hypothesized that “bottom-up” information processing of “content” elicits local, high frequency (beta-gamma) oscillations, whereas “top-down” processing is “contextual”, characterized by large scale integration spanning distant cortical regions, and implemented by slower frequency (theta-alpha) oscillations. This corresponds to a mechanism of cortical information transactions, where synchronization of beta-gamma oscillations between distant cortical regions is mediated by widespread theta-alpha oscillations. It is the aim of this paper to express this hypothesis quantitatively, in terms of a model that will allow testing this type of information transaction mechanism. The basic methodology used here corresponds to statistical mediation analysis, originally developed by (Baron and Kenny 1986). We generalize the classical mediator model to the case of multivariate complex-valued data, consisting of the discrete Fourier transform coefficients of signals of electric neuronal activity, at different frequencies, and at different cortical locations. The “mediation effect” is quantified here in a novel way, as the product of “dual frequency RV-coupling coefficients”, that were introduced in (Pascual-Marqui et al 2016, http://arxiv.org/abs/1603.05343). Relevant statistical procedures are presented for testing the cross-frequency mediation mechanism in general, and in particular for testing the von Stein & Sarnthein hypothesis.