Detrending Moving Average, Power Spectral Density, and Coherence: Three EEG-Based Methods to Assess Emotion Irradiation during Facial Perception

Abstract

Understanding brain reactions to facial expressions can help in explaining emotion-processing and memory mechanisms. The purpose of this research is to examine the dynamics of electrical brain activity caused by visual emotional stimuli. The focus is on detecting changes in cognitive mechanisms produced by negative, positive, and neutral expressions on human faces. Three methods were used to study brain reactions: power spectral density, detrending moving average (DMA), and coherence analysis. Using electroencephalogram (EEG) recordings from 48 subjects while presenting facial image stimuli from the International Affective Picture System, the topographic representation of the evoked responses was acquired and evaluated to disclose the specific EEG-based activity patterns in the cortex. The theta and beta systems are two key cognitive systems of the brain that are activated differently on the basis of gender. The obtained results also demonstrate that the DMA method can provide information about the cortical networks’ functioning stability, so it can be coupled with more prevalent methods of EEG analysis.

Method for Assessing the Influence of Phobic Stimuli in Virtual Simulators

“…A coherence matrix MC with the size of CH × CH is constructed, for which the average coherence in the average frequency range is calculated for all combinations of channels. Based on the approach used in [58], the upper triangle matrix is formed with nonzero elements equal to the coherence between channels that are statistically reliable. The validity of the coherence coefficient between pairs of electrodes is addressed by conducting surrogate data analysis [59].…”

Research on background harmonic portrait technology

“…In practice, due to the lack of prior knowledge of the background harmonic voltage pattern, the supervised analysis method is not suitable for identification [6]. As an unsupervised analysis method, clustering can divide data into several categories according to their similarity, so it can be used for pattern recognition of background harmonic voltage [7].…”