Effect of Artifacts on the Interpretation of EEG-Based Functional Connectivity Estimation Using Partial Directed Coherence

The proposed work investigates the design of an algorithm to enhance the functional connectivity for human brain using electroencephalogram (EEG) signals. The partial directed coherence (PDC) approach was used to estimate the right analysis of the functional connectivity. However, the presence of artifacts due to eye movement/blinking and muscles movement affects the correct estimation of functional connectivity. Therefore, the time-average differencing (TAD) method was developed to remove the artifacts. Two methods based on discrete wavelet transform (DWT) denoising and moving average filter are used, and comparison is made to access their effectiveness. Between the two methods, the TAD MA method removes all artifacts and ensures correct estimation of functional connectivity using PDC.