A brain signal-based credibility assessment approach

Deception detection is important for legal, moral and clinical purposes but still it is harder even for security officers and judges. Therefore an effective,light weight approach is a must.There are several technologies used in deception detection. EEG based deception detection is one such approach. P300 wave is most commonly used in EEG based deception detection which depends on a stimuli. The study provides an alternative approach to deception detection instead of using P300.Twelve subjects were participated to the study and EEG signals were recorded while they were telling truths and lies. The preprocessed EEG data then fed in to feature extraction and machine learning algorithm alone with Common Spatial Patterns (CSP) paradigm to create a model. Logistic regression classifier was used as the machine learning algorithm to classify the eeg signal. The test data were used on the trained model with cross validation. There were significant difference between truth telling and lying signals. The average rate of correctly predicted the class was 76%.