EEG/PPG effective connectivity fusion for analyzing deception in interview

Ali Motie-Nasrabadi | Marzieh Daneshi Kohan | Mohammad Bagher Shamsollahi | Ali Sharifi

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=148122
Date :  2020/02/14
Publish in :    Signal, Image and Video Processing
DOI :  https://doi.org/https://doi.org/10.1007/s11760-019-01622-1
Link :  http://dx.doi.org/https://doi.org/10.1007/s11760-019-01622-1
Keywords :Electroencephalogram,Deception detection, photoplethysmogram, Effective connectivity, Wavelet, Classification

Abstract :

---

In this research, the interaction between electroencephalogram (EEG) and, a cardiac parameter, photoplethysmogram (PPG), using connectivity measures to emphasize the importance of autonomic nervous system over the central nervous system during a deception is investigated. In this survey, connectivity analysis was applied, since it can provide information flow of brain regions; moreover, lying and truth appear to be cohered with the flow of information in the brain. Initially, a new wavelet-based approach for EEG/PPG effective connectivity fusion was introduced; then, it was validated for 41 subjects. For each subject, after extracting specific wavelet component of EEG and PPG signals, an effective connectivity network was generated by a generalized partial direct coherence and a direct directed transfer function. The results showed that grand average connectivity patterns were different in some regions for guilty and innocent subjects. The classification results demonstrated that lying could be discriminated from truth with the average accuracy of 84.14 by the leave-one-subject-out method. The present results contribute new information about coupling between EEG and PPG signals.