Effective Connectivity in Cortical Networks During Deception: A Lie Detection Study Based on EEG

Abstract

Previous studies have identified activated regions associated with deceptive tasks and most of them utilized time, frequency, or temporal features to determine deceptive responses. However, when deception behaviors occur, the functional connectivity pattern and the communication between different brain areas remain largely unclear. In this study, we explored the most important information flows between different brain cortices during deception. First, we employed the guilty knowledge test protocol and recorded on 64 electrodes' electroencephalogram (EEG) signals from 30 subjects (15 guilty and 15 innocent). EEG source estimation was then performed to compute the cortical activities on the 24 regions of interest (ROIs). Next, effective connectivity was calculated by partial directed coherence (PDC) analysis applied to the cortical signals. Furthermore, based on the graph theoretical analysis, the network parameters with significant differences were extracted as features to identify two groups of subjects. In addition, the ROIs frequently involved in the above network parameters were selected and based on the difference in the group mean of PDC values of all the edges connected with the selected ROIs, we presented the strongest information flows (MIIF) in the guilty group relative to the innocent group. Experimental results first show that the optimal classification features are mainly indegree and out-degree measures of the ROI and the high classification accuracy for four bands demonstrated that the proposed method is suitable for lie detection. In addition, the frontal-parietal network was found to be most prominent among all the MIIFs in four bands. Finally, combining the neurophysiology signification of four frequency bands, respectively, we analyzed the roles of all the important information flows to uncover the underlying cognitive processes and mechanisms used in deception.

Exploring the Functional Brain Network of Deception in Source-Level EEG via Partial Mutual Information

“…Other studies have found that enhanced communication between the 44R and 45L brain regions is associated with an increased working memory load during the act of lying [14,58]. The 44R region has been identified as a key "hub" in connectivity networks, potentially reflecting brain activity related to episodic memory and new information encoding [59]. The BA 9R/10R regions have been linked to strategic processes in memory recall, recognition, and various executive functions, and the BA44 region has been implicated in selective response suppression in go/no-go tasks [60].…”

Temporal and spatial variability of dynamic microstate brain network in disorders of consciousness

“…After applying the first step of filter‐based feature selection, an OFS was created using a wrapper strategy (see Figure 3B) for feature selection 29,30 . First, TFS in each microstate was ranked based on mRMR, and the top‐ranked new features were gradually added to the feature space to form a new feature vector.…”

“…After applying the first step of filter‐based feature selection, an OFS was created using a wrapper strategy (see Figure 3B ) for feature selection. 29 , 30 First, TFS in each microstate was ranked based on mRMR, and the top‐ranked new features were gradually added to the feature space to form a new feature vector. Then, a five‐fold cross‐validation (CV) strategy was applied to the ranked TFS, with the samples randomly divided into a training dataset (Dtra) and a testing dataset (Dtes) in a 4:1 ratio.…”

“…After applying the first step of filter-based feature selection, an OFS was created using a wrapper strategy (see Figure 3B) for feature selection. 29,30 First, TFS in each microstate was ranked based The MIFCN-Cortical connections corresponding to optimal feature subset. In addition, the present study used the mRMR method to find the set of features in the original feature set that had the greatest relevance to the final output, but the least relevance between the features.…”

Analysis of Weight-Directed Functional Brain Networks in the Deception State Based on EEG Signal

“…However, the results of Labkovsky et al [15] demonstrated that the false positive rate of deception detection utilizing P300 remains significant. Based on this, researchers used nonlinear analysis techniques to extract features from multi-channel ERP and machine learning to identify and classify features, which not only facilitates further decoding of the information processing mechanism of the brain during deception, but also further improves the accuracy of deception detection [2], [16], [17].…”