Efficacy of adaptive directed transfer function for neural connectivity estimation of EEG signal during meditation

Abstract

Electroencephalogram (EEG) is widely used in cognitive science, neuroscience and physiological research. It is a good mean to observe cognitive response that depends on time. EEG has many advantages over other techniques owing to its non-invasiveness, low cost and high temporal resolution. But one of the major challenges of EEG signal study is the huge data dimensionality which makes signal processing and subsequent analysis an uphill task. The aim of this study is to obtain a model for better neural connectivity analysis which illustrates meditation's dynamic mind-body response. Accordingly the EEG data is being collected during meditation (Kriya Yoga). In order to calculate and visualize the time-frequency representations of each electrode, a time varying Granger Causality based connectivity estimators named Directed Transfer Function (DTF) and adaptive DTF (ADTF) among all scalp electrodes have been computed in meditator group. The ADTF can be derived from the coefficients of a time-varying multivariate autoregressive (TVAR) model fitted to the data obtained during meditation. We define this time-varying measure of causality as the adaptive directed transfer function (ADTF) and compare its ability with the conventional DTF for meditator group. Both ADTF and Conventional DTF were calculated in meditator. The obtained simulation results of adaptive DTF and conventional DTF shows better neural connectivity and gives useful information in meditator group. However, to accomplish this task, surrogate data statistics has been used in both the mentioned models to validate the models. It was found that the ADTF has the capability to distinguish the dynamic changes in the primary source of the information outflow. The results obtained both by using ADTF and conventional DTF method were compared in meditator group subsequently.