Home › Polygraph Research › Machine Learning
Subject index
Machine Learning
This topic encompasses 397 peer-reviewed studies, primarily in the field of neuroimaging & new technologies. Research in this area spans from the 0s to the 2020s.
397Studies
0s–2020sSpan
13Categories
1040Researchers
Research categories
Key researchers
Studies tagged “Machine Learning” [397]
All 397 peer-reviewed studies in this research area, sorted by year.
[001]
(2026)
Large Language Models Simulating Deception and Coalition in Social Deduction Game
LLM agents demonstrated sophisticated deception strategies including agreeableness tactics and nomination mimicry to blend with truthful players, though they exhibited limitations in subtle deception without… [002] (2026)Lie Detection Using Facial Expressions and Speech Analysis
Lecture Notes in Networks and Systems The multimodal integration of facial and speech analysis demonstrates superior performance compared to single-modality approaches, leveraging complementary behavioral cues including micro-expressions, eye movements, pitch variation,… [003] (2026)Enhanced visualisation of concealed target objects by infrared thermography and machine learning
Infrared Physics & Technology The Fuzzy-c clustering approach outperformed K-means clustering in localizing concealed objects and successfully eliminated false signals from residual palm heat, demonstrating the potential of machine… [004] (2026)Buyer–Seller-Deception-Game Dataset: A new comprehensive dataset for facial expression based deception detection in economic contexts
Intelligent Systems with Applications OpenFace-based AU features performed best on clean, controlled recordings while CNN-based predictors outperformed on challenging in-the-wild data with lower video quality. Multimodal approaches combining visual… [005] (2026)A novel method based on variational mode decomposition for lie detection
Egyptian Informatics Journal VMD-based signal decomposition significantly enhanced P300 detection clarity and stability during deception compared to classical filtering methods, improving the reliability of EEG-based lie detection through… [006] (2026)Deception Detection: Using Machine Learning to Analyze 911 Calls
Personality and Social Psychology Bulletin The random forest model achieved 68.2% overall accuracy with 68.7% sensitivity and 67.7% specificity. Performance varied by context: homicide reports showed 71.2% accuracy with 77.3%… [007] (2026)Neural Signatures of Deception: An Explainable Machine LearningApproach Using EEG Signals
The LGBM-based framework achieved 97.7% accuracy in classifying truthful versus deceptive EEG responses, significantly outperforming conventional machine learning classifiers while maintaining model interpretability through feature… [008] (2025)Cognitive Computing Frameworks for Scalable Deception Detection in Textual Data
Big Data and Cognitive Computing CoTAM with sentence decomposition achieved statistically significant improvements (p < 0.05) over baseline methods. LLaMA-3-8B reached F1_lie of 0.35 compared to Gemma-3-27B's 0.19 in one-shot… [009] (2025)Lie Detection System using Multimodal Biometric Analysis
Journal of Information Systems Engineering and Management The system demonstrated qualitative feasibility of multimodal data fusion, showing that physiological arousal patterns combined with facial expression analysis could be integrated for lie detection,… [010] (2025)A Lightweight Visual Deception Detection Method Using Dynamic Images and a Customized SmallCNN–BiLSTM Architecture
2025 17th International Conference on Knowledge and System Engineering (KSE) The SmallCNN-BiLSTM architecture with dynamic images achieves competitive deception detection accuracy while offering substantial computational efficiency gains, making it suitable for real-world deployment in resource-constrained… [011] (2025)Neurocognitive Lie Detection Using EEG: Fast-Converging Deep Learning Models for Innocent and Guilty Classification in Neurological Disorders
The framework achieved accuracy above 80% with sensitivity and specificity exceeding 75% while converging faster than conventional classifiers, demonstrating that optimized deep learning can maintain… [012] (2025)Artificial Intelligence in Witness Credibility Assessment: The Role of Biometrics, Voice Analytics and Machine Learning in Judicial Cross-Examination
Journal of Asian Development Studies Neural Networks achieved 97% accuracy and 98% recall in witness credibility assessment, with 98% true positive rate, significantly outperforming voice analytics (92%) and other models,… [013] (2025)The Artificial Intelligence Application Concept in the System of Polygraph-Based Psychophysiological Lie Detection
Juridical Psychology The paper establishes a comprehensive conceptual framework defining AI application in polygraph systems, identifying implementation mechanisms, guiding principles, and operational forms, while systematically analyzing advantages,… [014] (2025)Effective faking of verbal deception detection with target-aligned adversarial attacks
Legal and Criminological Psychology Target-aligned adversarial attacks using GPT-4 reduced both human and machine deception detection performance to chance levels (d = -0.07 to -0.04 for humans, 51% accuracy… [015] (2025)Unmasking Deception: Examining Receivers' Emotional Responses with Automated Facial Expression Analysis in Veracity and Valence
Emotional mimicry was significantly lower for deceptive messages (F(1,505) = 32.64, p < .001, η²p = 0.06), with the strongest effect for positively framed deceptions.… [016] (2025)Tetralet Attention Enabled Modified N-Adam Optimised Distributed Capsule Network for Lie Detection from Electroencephalogram Signals
International Journal of Biomedical Engineering and Technology Specific results unavailable due to inaccessibility of the full paper text. [017] (2025)Evaluation of sampling rate effect on deception detection performance with audio features
AIP Conference Proceedings The study identifies optimal audio sampling rates for deception detection from speech, quantifying the tradeoffs between acoustic feature quality and computational efficiency to inform practical… [018] (2025)Multi-source Multimodal Progressive Domain Adaption for Audio-Visual Deception Detection
Proceedings of the 1st International Workshop & Challenge on Subtle Visual Computing The MMPDA framework achieved 60.43% accuracy and 56.99% F1-score on cross-domain deception detection, surpassing the 1st place team by 5.59% on F1-score and demonstrating effective… [019] (2025)Facial expression-based lie detection: A survey of micro expression analysis, datasets and challenges
AIP Conference Proceedings The survey finds that automated micro-expression analysis shows promise for deception detection, with machine learning methods outperforming human judgments, though accuracy varies widely (63-78%) depending… [020] (2025)Neurophysiological Approaches to Lie Detection: A Systematic Review
Brain Sciences The Concealed Information Test with P300 was the most frequently employed protocol. Bandpass filtering and Discrete Wavelet Transform emerged as preferred preprocessing and feature extraction… [021] (2025)Variance in Veracity: Comparing Two Neuro-Paradigms for CNV Analysis Applied to Deception Detection
2025 2nd Global AI Summit - International Conference on Artificial Intelligence and Emerging Technology (AI Summit) NP2 paradigm achieved highest classification accuracy of 82.2% using Gradient Boosting in visual modality, with stronger CNV responses indicating that embedding deception among familiar stimuli… [022] (2025)Lie Detection in Social VR Using Multimodal Data
2025 IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW) Gaze-based features alone achieved 87% accuracy in detecting lies within social VR environments using random forest classification, demonstrating the feasibility of passive deception detection using… [023] (2025)High-Risk AI Systems—Lie Detection Application
Future Internet AI lie detection systems at borders qualify as high-risk under EU AI Act requirements. Analysis of iBorderCtrl revealed 75% claimed accuracy but 24.45% false positive… [024] (2025)Exploring Vision-Based Features for Detecting Deception in Well-Being: A Cross-Domain Comparison
2025 IEEE 19th International Conference on Automatic Face and Gesture Recognition (FG) Well-being deception showed fewer and more localized visual features concentrated around the nose ridge compared to biography and academic domains. Joy-related facial action units (AU… [025] (2025)A Review on Robustness of Deception Detection System Based on Eye Movement Features
2025 11th International HCI and UX Conference in Indonesia (CHIuXiD) Most eye movement-based deception detection systems achieved only low to medium robustness scores. Multimodal approaches demonstrated higher robustness than unimodal approaches, but challenges including limited… [026] (2025)DecepTech AI: Advancing Lie Detection Through Multimodal Nonverbal Behavior Analysis
Lecture Notes in Networks and Systems The system successfully integrated multiple nonverbal behavioral channels using machine learning to detect deception patterns, demonstrating capability for real-time processing and avoiding the intrusiveness of… [027] (2025)Recognition of emotions by analysing facial expressions with FaceReader (Noldus) vs detection of deception by polygraph examination: A pilot study
European Polygraph The study explored the integration of facial expression analysis technology with traditional polygraph examination to identify emotions accompanying deceptive responses during concealed information testing. [028] (2025)You really don't recognise him? The eye-tracker as a forensic tool for concealed knowledge detection
European Polygraph The study explored whether eye-tracking combined with deep learning can objectively determine recognition of individuals even when examinees deny familiarity, representing a potential advancement in… [029] (2025)Explainable AI for Unraveling the Significance of Visual Cues in High Stakes Deception Detection
IEEE Access Eyebrow movements emerged as the most reliable visual indicator of deception in high-stakes courtroom settings. The Multi-layer Perceptron model achieved 88% accuracy and 92.86% recall… [030] (2025)Comparing OpenFace and Deep Learning Models for Deception Detection in Video Calls
2025 14th International Symposium on Image and Signal Processing and Analysis (ISPA) CNN-based AU detection outperforms OpenFace on challenging Real-life Trial data with lower video quality and unstable head pose, while OpenFace performs better in controlled settings… [031] (2025)An Efficient Lie Detection with Movement Tracks of Larynx and Eye Flutters During Interrogation Using Machine Learning and Deep-Computing
Wireless Personal Communications The system successfully automated detection of micro-level eye and laryngeal movements invisible to the naked eye, demonstrating that non-verbal signals including blink patterns and swallowing… [032] (2025)Efficient Deception Detection Using Multimodal Facial and Audio Transcript Features
International Journal of Scientific Research in Science and Technology The Modified CNN model achieved 91.2% accuracy using multimodal fusion of facial action units, eye gaze, and audio prosodic features with attention-based feature fusion, outperforming… [033] (2025)Detecting Deepfakes Through Emotion?: Facial Expression and Emotional Contagion as Dual Indicators of Deepfake Credibility
Applied Cognitive Psychology Deepfakes exhibited lower overall and negative emotions compared to authentic videos. Counterintuitively, audiences' reduced emotional responses to deepfakes were associated with higher perceived credibility, mediated… [034] (2025)A Comprehensive Review and Future Prospects of Lie Detection Using Machine Learning
Communications in Computer and Information Science Full text not accessible; unable to extract specific findings from the review. [035] (2025)Artificial intelligence in lie detection: Why do cognitive theories matter?
New Ideas in Psychology Mechanism-driven theories enable researchers to identify what features AI extracts and how those features relate to deception, allowing forensic experts to anticipate error patterns and… [036] (2025)A Novel Approach for Enhancing Audio Modality in Multimodal Deception Detection Environment Using Liquid Neural Networks
2025 15th International Conference on Electrical Engineering (ICEENG) LNNs achieved 65.33% accuracy on DOLOS dataset and 87.5% on courtroom dataset, representing 5.04% and 8.5% improvements over previous audio-based benchmarks through dynamic adaptation to… [037] (2025)@LM DeceptionNet: A multimodal approach for efficient transfer learning-based deception detection
Knowledge-Based Systems The @LM DeceptionNet framework demonstrated consistent cross-scenario generalization with enhanced accuracy and robustness through adaptive multimodal fusion, achieving better computational efficiency with fewer parameters compared… [038] (2025)Truth Under Pressure: A Deep Learning-Based Lie Detection System for Online Lending Using Voice Stress and Response Latency
International Journal of Advanced Computer Science and Applications The lie detection model achieved 70% testing accuracy with 70.9% F1-Score when analyzing voice stress and response latency from court testimony videos, while the Llama… [039] (2025 • DEEP LEARNING)Development of a Deep-Learning-Based Computerized Scoring Algorithm for Polygraph Data
MDPI Journals Developed a Korean polygraph scoring system using deep neural networks. DNN-based algorithm outperformed both PolyScore and OSS-3 on test data by accounting for bio-signal nonlinearity. [040] (2025)AN ENHANCED ALGORITHM TO IMPROVE THE ACCURACY OF LIE DETECTION SYSTEM BASED ON EEG SIGNAL
Kufa Journal of Engineering The enhanced signal processing pipeline achieved 86% accuracy on benchmark data and 78% accuracy on newly collected participant data, with a 23.2% improvement in signal-to-noise… [041] (2025)Multimodal Deception Detection Using Linguistic and Acoustic Features
Machine Learning, Deep Learning and AI for Cybersecurity A single-layer BiLSTM model achieved 93.57% accuracy and 94.48% F1 score on textual data from court testimony transcripts. ResNet50 performed best on audio data, and… [042] (2025)Lie Detection Using Audio Classification
Lecture Notes in Electrical Engineering The CNN-based audio classification model achieved 93.5% accuracy in detecting deception from courtroom audio data, demonstrating the effectiveness of deep learning for automatic feature extraction… [043] (2025)A psycholinguistic NLP framework for forensic text analysis of deception and emotion
Frontiers in Artificial Intelligence The researchers successfully identified both guilty parties from 18 suspects using combined entity correlation, deception scoring via Empath, emotion analysis through Hugging Face transformers, and… [044] (2025)A multilayer perceptron neural network prediction approach to polygraph scoring
INTERNATIONAL JOURNAL OF COMPUTERS COMMUNICATIONS & CONTROL The MLP neural network achieved 80% correct classification rate predicting deception levels in 400 offender examinations based on four psychophysiological indicators, with performance consistent with… [045] (2025)Real-time cross-cultural lie detection system via multimodal fusion: microexpression enhancement and adversarial defense for forensics
Advances in Engineering Innovation The MPBFS achieved 84.6% accuracy (F1=0.87) on cross-cultural datasets at 23 FPS, with a 31% improvement in Cultural Stability Index and 42% reduction in adversarial… [046] (2025)A Filtering Response Time Concealed Information Test for Searching for Relevant Concealed Items
Applied Cognitive Psychology The filtering RT-CIT protocol reduced testing time by approximately 59% (from 17 to 7 minutes) compared to conventional approaches while maintaining similar probe identification accuracy… [047] (2025)Text Based Lie Detection Using a Combination of BERT and CNN Models
2025 15th International Conference on Information & Communication Technology and System (ICTS) The optimal configuration achieved an F1-score of 0.9154 on the Deceptive Opinion Spam Dataset using BERT Base Uncased with fully unfrozen encoders, kernel sizes [2,3,4],… [048] (2025 • MULTIMODAL ML)Multimodal Machine Learning for Deception Detection Using Behavioral and Physiological Data
Scientific Reports (2025) Created multimodal dataset spanning EEG, ECG, EOG, eye-gaze, GSR, audio, and video from 100+ subjects. Demonstrated machine learning integration of multiple modalities for improved detection. [049] (2025)Multimodal Deception Detection via Cognitively Guided Inconsistency Modeling
Proceedings of the 1st International Workshop & Challenge on Subtle Visual Computing The cognitively-guided framework successfully captured subtle, asymmetric, and ambiguous deceptive patterns by modeling modality leakage through selective channel replacement and aligning prediction confidence with psychological… [050] (2025)Effects of Soft-Domain Transfer and Named Entity Information on Deception Detection
Lecture Notes in Computer Science The study found improvements in accuracy over baseline models, with named entity information augmentation producing notable improvements of up to 11.2%. Jensen-Shannon distance correlated moderately… [051] (2025)Neuroscientific and Biometric Methodologies in Forensic Contexts: A Critical Review
Psychology and Behavioral Sciences The review provides evidence-based evaluation of cutting-edge forensic tools including fMRI, ERP, brain fingerprinting, and machine learning techniques that may complement or supplement traditional investigative… [052] (2025)Exploring Audio-Visual Fusion Methods in Foundation Model-Based Deception Detection
2025 Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) Audio-visual fusion using Q-Former architecture with foundation models consistently outperformed unimodal approaches, with systems aggregating embeddings from all layers achieving 3.78% accuracy improvements over last-layer-only… [053] (2025)A Multimodal Polygraph Framework with Optimized Machine Learning for Robust Deception Detection
MDPI Journals The study addresses critical gaps by providing an open multimodal dataset and demonstrating improved generalization through PSO-enhanced machine learning approaches to physiological data fusion. [054] (2025)Teaching AI Agents to Lie at Social Deduction Games
Game AI Uncovered The system successfully enabled AI agents to deceive human players and detect deception by maintaining probabilistic worldviews for each player, evaluating claim consistency against belief… [055] (2024)Lie Detection Based on Acoustic Analysis
Journal of Voice Voice Onset Time achieved 88.8% accuracy (both sensitivity and specificity) in detecting deception, with 95% confidence intervals of 0.803-0.973, demonstrating that acoustic features can effectively… [056] (2024)Deception Indicators Prediction for Lie Detection in Dialogues
2024 16th International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC) Full paper not accessible for analysis. [057] (2024)Detecting deception with artificial intelligence: promises and perils
Trends in Cognitive Sciences AI-based deception detection systems lack sufficient scientific validation and may repeat the mistakes of polygraph technology, potentially leading to systematic discrimination against innocent individuals. [058] (2024)EEG-based deception detection using weighted dual perspective visibility graph analysis
Cognitive Neurodynamics The WDPVG method achieved 86.25% accuracy with SVM classification, with frontal and parietal lobe activity showing the strongest discrimination between truth-telling and deception through graph-based… [059] (2024)Application of Artificial Neural Network in Deception Detection Based on Visual Cues
IFMBE Proceedings The LRCN (Long-term Recurrent Convolutional Network) model achieved the highest classification accuracy of just over 70%, outperforming feature-based LSTM approaches that ranged from 61-70% accuracy… [060] (2024)Deception detection using machine learning and deep learning techniques: A systematic review
Natural Language Processing Journal The review identifies that human deception detection accuracy remains at only 54%, barely better than chance, while modern facial expression-based ML systems offer non-intrusive, cost-effective… [061] (2024)FTM: The Face Truth Machine—Hand-crafted features from micro-expressions to support lie detection
Computer Vision and Image Understanding FTM achieved satisfying cross-dataset transferability and can analyze facial expressions across a wide range of poses. The use of hand-crafted features enables interpretation of which… [062] (2024)(When) Would You Lie to a Voicebot?
SSRN Electronic Journal Participants were more likely to lie when hearing a feminine voice than a masculine voice, and more likely to lie to a feminine-voiced basic voicebot… [063] (2024)The Most Important Features of Lie Detection Using Voice Stress
Al-Iraqia Journal for Scientific Engineering Research The Random Forest classifier achieved 88% accuracy in distinguishing deceptive from truthful courtroom testimony using only acoustic features, with deceptive speech exhibiting higher levels of… [064] (2024)Assessing partial errors via analog gaming keyboards in response conflict tasks: A proof-of-concept study with the concealed information test
Behavior Research Methods Probe trials showed 2.93% partial errors versus 0.46% for irrelevant items, with strong statistical evidence (Bayes Factor = 275) supporting the detection of response conflict… [065] (2024)Analysis, Evaluation, and Future Directions on Multimodal Deception Detection
Technologies The review found that 85% of studies used facial features, 50% audio, and 38% textual features; 90% employed model-independent fusion with late fusion being most… [066] (2024)Experimental economics for machine learning—a methodological contribution on lie detection
PLOS ONE Machine learning algorithms trained on video data from an economic dice-rolling experiment achieved 67% lie detection accuracy when ground truth was verified through experimental design,… [067] (2024)MVis4LD: Multimodal Visual Interactive System for Lie Detection
Lecture Notes in Computer Science The research successfully developed a web-based multimodal interactive system that enables users to explore deception detection data through natural language, demonstrating the feasibility of integrating… [068] (2024)Flexible-Modal Deception Detection with Audio-Visual Adapter
2024 IEEE International Joint Conference on Biometrics (IJCB) The AVA framework achieved superior performance compared to other multi-modal fusion methods under flexible-modal settings, with audio modality generally outperforming vision-only approaches, while maintaining robust… [069] (2024)Blue Sky: Multilingual, Multimodal Domain Independent Deception Detection
Proceedings of the 2024 SIAM International Conference on Data Mining (SDM) The paper identifies that universal linguistic cues to deception in non-English languages remain unknown, low-resource language deception detection lacks sufficient labeled data, and multimodal deception… [070] (2024)LIE DETECTION USING FACIAL MICRO-EXPRESSION FEATURES WITH MACHINE LEARNING
MINAR International Journal of Applied Sciences and Technology The Sobel filter outperformed Canny edge detection, and mouth/nose regions proved more discriminative than eyes for lie detection, achieving F1-scores of 0.996 (truth) and 0.990… [071] (2024)Domain-Agnostic Adapter Architecture for Deception Detection: Extensive Evaluations with the DIFrauD Benchmark
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) The domain-agnostic adapter architecture achieved state-of-the-art performance on the DIFrauD benchmark across seven fraud domains, consistently outperforming all competing methods and revealing that representation quality… [072] (2024)An Automatic Lie Detection Model Using EEG Signals Based on the Combination of Type 2 Fuzzy Sets and Deep Graph Convolutional Networks
Sensors The proposed model achieved >95% accuracy, precision, sensitivity, and specificity for distinguishing truths from lies, with remarkable robustness maintaining >90% accuracy even in extremely noisy… [073] (2024)The Good, the Bad, and the Ugly: The Role of Ai Quality Disclosure in Lie Detection
When relying on low-quality AI advisors without disclosure, participants' truth-detection rates fell below their unassisted baseline (approximately 36%), recovering only when AI effectiveness was revealed.… [074] (2024)Analysis of Deception Detection Databases Using Mathematical Statistics
Applied Mathematics & Information Sciences The study successfully characterized eight deception detection databases and developed a novel multimodal dataset with rigorous data collection protocols, including MFCC audio analysis, facial feature… [075] (2024)Physiological Reactions Profiling in Polygraph Testing: Insights from Fuzzy C-Means Clustering Analysis
Acta Polytechnica Hungarica Fuzzy clustering identified distinct physiological profiles across 18 parameters with a moderate R² of 0.371 and a low silhouette coefficient of 0.010, indicating that physiological… [076] (2024)Deception Detection Algorithm Based on Global and Local Feature Fusion with Multi-head Attention
2024 3rd International Conference on Image Processing and Media Computing (ICIPMC) The study demonstrates that fusion of local facial features (AU strength and gaze angle) with global temporal patterns through Siamese Transformer architecture and multi-head attention… [077] (2024)Multi-Modal Deception Detection Using Deep Learning
2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT) The multimodal deep learning system achieved 94.23% overall accuracy in detecting deception from trial hearing recordings, substantially exceeding previous benchmarks and human detection capabilities (62%… [078] (2024)Deception Detection from Linguistic and Physiological Data Streams Using Bimodal Convolutional Neural Networks
2024 5th International Conference on Information Science, Parallel and Distributed Systems (ISPDS) The bimodal CNN successfully integrated physiological measurements (heart rate, skin conductance, respiration, skin temperature) with linguistic features from transcripts, demonstrating that neural networks can effectively… [079] (2024)Combining recognition, conflict‐monitoring and feedback‐related ERPs to detect concealed autobiographical information
Psychophysiology Three ERP indicators (N200, recognition P300, and feedback P300) significantly discriminated guilty from innocent participants, with combined classification achieving AUC = 0.91. The FRN did… [080] (2024)Video-Based Deception Detection and Financial Fraud
SSRN Electronic Journal Visual deception scores derived from IPO roadshow videos significantly predict financial fraud. The effect is strongest for deception cues related to financial performance discussions, and… [081] (2024)Training on dirty labels: Rejoinder to Kotsoglou and Biedermann
Forensic Science International: Synergy The authors defend their ML approach as practically useful for quality assurance despite training on potentially fallible human labels, arguing that polygraph use remains legitimate… [082] (2024)The Walk of Guilt: Multimodal Deception Detection from Nonverbal Motion Behaviour
Information Classification accuracy reached up to 88% through hybrid feature fusion, with visual signals showing highest single-modality performance followed by accelerometer and acoustic signals, averaging 60%… [083] (2024)Statistical Analysis of Video Databases for Deception Detection Tasks
Applied Mathematics & Information Sciences The study identified significant limitations in publicly available deception detection databases and presented a novel dataset incorporating video, audio, and eye gaze data with systematic… [084] (2024)Verification of using radar vital sign monitors in deception tests (2023)
Applied and Computational Engineering The study produced conclusive findings demonstrating that radar technology can non-invasively extract vital signs during deception tests with validated accuracy, enabling more flexible and less… [085] (2024)Slow Tone: Detecting White Lie Disclosures Using Response Latency
Positive management tone following high response latency predicts immediate positive stock price reactions that reverse over time, with effects most pronounced in firms with lower… [086] (2024)Automatic Detection of Verbal Deception in Romanian With Artificial Intelligence Methods
Studia Universitatis Babeș-Bolyai Informatica Results showed that for topic-specific datasets with a common subject, general text representations such as TF-IDF and LSA outperformed psycholinguistic LIWC categories in detecting deception,… [087] (2024)A semi-supervised high-quality pseudo labels algorithm based on multi-constraint optimization for speech deception detection
Computer Speech & Language The HQPL-MC algorithm successfully improved deception detection performance through multi-constraint optimization of pseudo-labels, combining noise-augmented feature extraction, autoencoder-based dimensionality reduction, and threshold-based quality filtering to… [088] (2024)Neuro-Linguistic Programming as a Verbal Lie Detection Tool: A Case Study of Pakistani Digital Media Programs
Kurdish Studies The RoBERTa-base model achieved 75% effectiveness for neurolinguistic deception detection but successfully identified only truth statements, not lies. Lie statements frequently comprised compound, complex, and… [089] (2024)Polygraph-based deception detection and Machine Learning: Combining the Worst of Both Worlds?
Forensic Science International: Synergy The study concludes that applying machine learning to inherently problematic polygraph data does not address the fundamental scientific validity issues and may create additional concerns… [090] (2024)Automatic deceit detection through multimodal analysis of High-Stake Court-Trials
IEEE Transactions on Affective Computing The proposed architecture improved state-of-the-art performance on the Real-Life Trial dataset, with all classifiers achieving better results than measured human performance for deception detection. [091] (2024)Detecting Deception in Natural Environments Using Incremental Transfer Learning
International Conference on Multimodel Interaction Incremental transfer learning with logistic regression achieved 80% deception detection accuracy. Electrodermal activity and skin temperature showed consistent significant differences between deceptive and truthful states… [092] (2024)How Entangled is Factuality and Deception in German?
Findings of the Association for Computational Linguistics: EMNLP 2024 The study found zero correlation between established linguistic deception cues and actual deception in German texts when factuality and belief are properly disentangled. Both traditional… [093] (2024)Unmasking Lies: A Literature Review on Facial Expressions and Machine Learning for Deception Detection
Procedia Computer Science Machine learning models achieve superior performance compared to human judges (AI: 72-97% accuracy vs. humans: 57%), with multimodal approaches outperforming single-modality systems. However, significant limitations… [094] (2024)Evaluating Physiological Indicators in Detecting Deception with an Automated Interviewing System
bioRxiv (preprint) The study represents one of the few systematic comparisons of multiple physiological modalities within a single experimental framework, conducted with a diverse Asian population to… [095] (2024)Nothing but lies: improving the validity of neural predictors of deception
Neural predictors trained to detect deception also predicted selfish nondeceptive choices, confirming a century-old validity concern. A novel dual-goal tuning machine learning method successfully removed… [096] (2024)A Review of Automatic Lie Detection from Facial Features
Journal of Nonverbal Behavior Automated facial lie detection systems achieve average accuracy of 61.87% to 72.93% depending on classifier type, based on average datasets of 52 individuals. Most studies… [097] (2024)From Polygraphs to Truth Machines: Artificial Intelligence in Lie Detection
Critical Humanities AI-enhanced lie detection systems introduce novel concerns around 'biomarkers of deceit,' algorithmic opacity, invasive data collection, and bias, potentially enabling pervasive 'truth machines' that displace… [098] (2024)TrueWatch: Polygraph Examination With Smartwatches
IEEE Transactions on Information Forensics and Security Modern wearable smartwatches can detect lies in the Guilty Knowledge Test protocol, though their detection ability is comparatively weaker than conventional polygraph instruments. [099] (2024)Detecting Deception Using MCI in Twitter
International Journal of Forensic Sciences Computer-based linguistic analysis of MCI-derived Twitter statements achieved 79% classification accuracy in distinguishing truthful reports, denials, and false claims—substantially outperforming human judgment at 54%. Speech… [100] (2024)A Tutorial for Deception Detection Analysis or: How I Learned to Stop Aggregating Veracity Judgments and Embraced Signal Detection Theory Mixed Models
Journal of Nonverbal Behavior The SDT-GLMM approach avoids data aggregations that conflate response bias with underlying sensitivity, outperforms traditional factorial methods, and provides distinct estimates for discriminability (d') and… [101] (2024)Automated Comparative Predictive Analysis of Deception Detection in Convicted Offenders Using Polygraph with Random Forest, Support Vector Machine, and Artificial Neural Network Models
Studies in Informatics and Control Machine learning models successfully predicted deception from polygraph physiological data, with clustering analyses revealing important individual differences in response patterns. The findings demonstrate potential for… [102] (2024)On the Impact of a Simulated Cognitive Augmentation to Detect Deception on Decision-making Confidence
ACM Symposium on Spatial User Interaction Visualization methods alone did not significantly affect decision-making confidence, but a strong positive correlation existed between users' comfort with cognitive augmentation technologies and their decision-making… [103] (2023)Deception detection with machine learning: A systematic review and statistical analysis
PLoS ONE Of 81 reviewed studies, detection accuracy ranged from 51% to 100%, with 19 works achieving above 90% accuracy. Monomodal approaches constituted 40.74% of studies, though… [104] (2023)Intelligent Eye-Tracker-Based Methods for Detection of Deception: A Survey
MDPI Journals Eye-tracking methods show promise as a contactless, non-invasive alternative to conventional polygraph examinations, with potential applications in mass-screening scenarios like airport security and pre-employment procedures. [105] (2023)Increased Interpretability for Model-Driven Deception: MARS LDRD Project
The project successfully increased interpretability of machine learning models for industrial control system physical processes, allowing cyber analysts and incident response personnel to more optimally… [106] (2023)Exploring Verbal Cues to Deception: Testing Quantitative Linguistic Methods on English and Spanish
Classification accuracy reached 93% for high-stakes police deception versus 71% for laboratory lies, with LIWC psycholinguistic features achieving 84.5% F-measure on Spanish corpus, demonstrating that… [107] (2023)Determining psycholinguistic features of deception in phishing messages
Information & Computer Security Words conveying certainty (e.g., always, never) and work-related features predicted higher end-user vulnerability to phishing, while achievement and reward language (e.g., earn, win, cash, money)… [108] (2023)Hierarchical Bayesian Modeling of Autonomic Responses during the Concealed Information Test
Japanese Journal of Physiological Psychology and Psychophysiology The hierarchical Bayesian state-space model successfully analyzes CIT physiological time-series data in natural measurement units without standardization, revealing typical response magnitudes and individual differences previously… [109] (2023)EEG Signal Classification for Concealed Information Test using Spider Monkey Candidate Rule Miner
Multimedia Tools and Applications The SM-CRM successfully generated an optimal rule set for binary classification of guilty versus innocent participants based on EEG data, balancing sensitivity, accuracy, and specificity… [110] (2023)Meta Learning Based Deception Detection from Speech
Applied Sciences CHAML with Wav2Vec 2.0 embeddings achieved 61.34% accuracy in person-independent deception detection, outperforming baseline approaches while using only a few labeled samples per test subject. [111] (2023)Optimization of Signal Processing Parameters in Psychophysiological Studies on the Example of GSR and PPG
Experimental Psychology (Russia) Universal signal processing parameters were validated that demonstrated superiority over traditional baseline-relative methods and proved applicable across various physiological signals including fMRI. [112] (2023)Deception Detection Deep Learning Comprehensive system Utilizing Explainable AI
2023 16th International Conference on Developments in eSystems Engineering (DeSE) The paper presents a conceptual framework integrating five sensing modalities with deep transfer learning and XAI techniques to create transparent, countermeasure-resistant deception detection systems. No… [113] (2023)Man vs machine – Detecting deception in online reviews
Journal of Business Research Automated text analysis methods show complementary strengths in detecting deceptive reviews, with incentivized reviews exhibiting linguistic patterns similar to unverified content. Multi-method integration analyzing quantity,… [114] (2023)Bayesian Generalized Linear Mixed Effects Models for Deception Detection Analyses
The SDT GLMM approach separates true discriminability (d-prime) from response bias (criterion), incorporates individual differences through random effects, and reveals substantially greater uncertainty in effect… [115] (2023)The influence of micro-expressions on deception detection
Multimedia Tools and Applications The custom-trained deep learning model achieved 74.17% accuracy in deception classification by detecting involuntary micro-expression changes, with shifts from neutral to happy expressions reliably indicating… [116] (2023)Hybrid Machine Learning Model for Lie-Detection
2023 IEEE 8th International Conference for Convergence in Technology (I2CT) The hybrid CNN-SVM model achieved approximately 58% accuracy in detecting deception from facial thermal images, demonstrating the technical feasibility of the approach but highlighting the… [117] (2023)A Context-Constrained Sentence Modeling for Deception Detection in Real Interrogation
INTERSPEECH 2023 The context-constrained model achieved 76.38% Unweighted Average Recall (UAR) at the local level and 73.15% UAR at the global level, significantly outperforming baseline methods that… [118] (2023)Intelligent techniques for deception detection: a survey and critical study
Soft Computing Machine learning and deep learning methods have demonstrated significant effectiveness in deception detection across multiple modalities, though current datasets lack diversity and ecological validity, being… [119] (2023)The Sound of Cooperation and Deception in High Stakes Interactions
Machine learning algorithms classified cooperative versus deceptive speech with 58-79% accuracy using only acoustic features, with higher pitch and lower intonation significantly associated with deceptive… [120] (2023)LieVis: A Visual Interactive Dashboard for Lie Detection Using Machine Learning and Deep Learning Techniques
2023 26th International Conference on Computer and Information Technology (ICCIT) BERT deep learning model, along with logistic regression and K-Nearest Neighbors, achieved 84% accuracy in detecting deception from courtroom testimony transcripts, outperforming traditional machine learning… [121] (2023)Hybrid Metaheuristics with Deep Learning Enabled Automated Deception Detection and Classification of Facial Expressions
Computers, Materials & Continua The study demonstrates a novel hybrid approach combining self-adaptive population-based firefly algorithm optimization with deep learning (MobileNet + LSTM) for facial expression-based deception classification in… [122] (2023)Counter-Deception Information Fusion Based on Reinforcement Learning Using the Guidance of Belief Entropy and Conflict Management
The reinforcement learning framework successfully detects deceptive evidence that provides little support to correct targets while supporting incorrect ones, outperforming previous conflict-based methods in special… [123] (2023)A Microcontroller-based Lie Detection System Leveraging Physiological Signals
2023 IEEE International Conference on Internet of Things and Intelligence Systems (IoTaIS) The Arduino-based system successfully captured real-time heart rate and GSR changes during questioning, demonstrating that low-cost microcontroller platforms can measure physiological responses associated with arousal,… [124] (2023)A Deep Learning Approach for Multimodal Deception Detection
Lecture Notes in Computer Science The multimodal deep learning approach achieved 96.14% accuracy and 0.98 ROC-AUC, significantly outperforming prior automated methods (77-92% AUC) and human performance (60-81% AUC). [125] (2023)Deep Learning Applications in Brain Computer Interface Based Lie Detection
2023 IEEE 13th Annual Computing and Communication Workshop and Conference (CCWC) Ensemble deep learning models demonstrated the most effective performance among tested architectures. Models showed asymmetric performance, with better accuracy detecting truths than lies. [126] (2023)Building a second-opinion tool for classical polygraph
Nature Scientific Reports The study successfully created a system that can catch mistakes in historical polygraph records by identifying examiner errors in interpreting raw physiological data. [127] (2023)Enhancing Deception Detection with Exclusive Visual Features using Deep Learning
International Journal of Performability Engineering Deep learning models using CNN architectures successfully identified deceptive behavior from visual features alone, with automated feature extraction eliminating the need for manual annotation of… [128] (2023)Speech Deception Detection Based on EMD and Temporal Neural Network
Computational Intelligence and Neuroscience The TCN-LSTM network achieved 86.2% accuracy and 86.0% F1 score on real-life trial dataset when combined with EMD preprocessing. EMD signal decomposition improved average accuracy… [129] (2023)Choosing to Lie: Strategic and Pathological Deception
Competition context produced significantly elevated deception compared to control conditions (Cohen's d > 1.0, p < 2.2e-16), with most participants showing no deception in neutral… [130] (2023)Automated Deception Detection from Videos: Using End-to-End Learning Based High-Level Features and Classification Approaches
arXiv preprint Facial expressions consistently outperformed gaze and head pose modalities, with multimodal fusion achieving AUC scores ranging from 0.65 in low-stakes scenarios to 0.86 in high-stakes… [131] (2023)Deception Detection using a Multimodal Stacked Bi-LSTM Model
2023 International Conference on Innovative Data Communication Technologies and Application (ICIDCA) The multimodal stacked Bi-LSTM model achieved significant improvements in deception detection compared to single-modality models, with multimodal approaches showing 10-15% accuracy improvements over unimodal methods. [132] (2023)Neural network applications in polygraph scoring — A scoping review
MDPI Journals Neural networks, particularly feedforward and convolutional architectures, show promise in improving polygraph accuracy over traditional methods, though further validation research is needed to determine optimal… [133] (2023)Deception Detection in Videos Using Robust Facial Features with Attention Feedback
Handbook of Dynamic Data Driven Applications Systems The Bayesian neural network model achieved 92.36% accuracy detecting deception from facial features alone, with an attention mechanism that identifies key behavioral moments and enables… [134] (2022)Detecting concealed information using functional near-infrared spectroscopy (fNIRS) combined with skin conductance, heart rate, and behavioral measures
Psychophysiology The combination of fNIRS, reaction time, skin conductance, and heart rate achieved an area under the ROC curve of 0.94, compared to individual measures ranging… [135] (2022)Fake Online Reviews: A Unified Detection Model Using Deception Theories
IEEE Access Non-verbal behavioral features outperformed verbal linguistic features in detecting fake reviews, and combining both feature types enhanced model performance. The theory-based model achieved high interpretability… [136] (2022)A Comprehensive Review on Features Extraction and Features Matching Techniques for Deception Detection
IEEE Access Review of 88 studies over 30 years reveals deception detection accuracies ranging from 50-85% across modalities, with vocal features outperforming visual cues. Critical limitations include… [137] (2022)Lying trolls: Detecting deception and text-based disinformation using machine learning
Cybersecurity and Cognitive Science Hybrid models combining disinformation training data with other cybercrime text datasets achieved detection accuracies of 60-100%, substantially outperforming singleton models and demonstrating cross-domain transferability of… [138] (2022)Implementation of Machine Learning in BCI Based Lie Detection
2022 IEEE World AI IoT Congress (AIIoT) Machine learning classification of multimodal BCI signals (EEG, fNIRS, HRV) achieved 75-79.4% accuracy in distinguishing truthful from deceptive responses, with truthful responses detected at 81.5%… [139] (2022)RCIT: An RSVP-Based Concealed Information Test Framework Using EEG Signals
IEEE Transactions on Cognitive and Developmental Systems The RCIT framework achieved 82.74% accuracy using linear discriminant analysis in subject-dependent testing and 64.07% accuracy in cross-subject testing. PZ and T7 electrode channels were… [140] (2022)Concealed Information Detection Using EEG for Lie Recognition by ERP P300 in Response to Visual Stimuli: a Review
WSEAS TRANSACTIONS ON INFORMATION SCIENCE AND APPLICATIONS The review found that P300-based CIT with EEG achieves accuracies of 81.9% to 91.83% across different methodologies, significantly outperforming traditional autonomic polygraph methods. Discrete Wavelet… [141] (2022)A Deep Learning Model for EEG-Based Lie Detection Test Using Spatial and Temporal Aspects
Computers, Materials & Continua The V-TAM model achieved 98.5% accuracy (p < 0.01) in lie detection, outperforming DL, DL-LSTM, and standard Attention models while requiring only 0.19 seconds per… [142] (2022)Deception Detection And Vulnerability Analysis Using A Multi-Level Clustering Machine Learning Algorithm In Business Transactions
SSRN Electronic Journal The multi-level clustering machine learning technique produced the greatest outcomes for fraud detection, while straightforward anomaly detection methods produced the poorest outcomes when evaluated on… [143] (2022)Looking for Lies: An Exploratory Analysis for Automated Detection of Deception
Word count-based machine learning models were most accurate at detecting deception (76% accuracy), with similar performance across video and text-based interviews, though significant gender bias… [144] (2022)Complex Paralinguistic Analysis of Speech: Predicting Gender, Emotions and Deception in a Hierarchical Framework
Interspeech 2022 The hierarchical framework achieved 88.4 ± 1.5% accuracy for deception detection, a 2.8% improvement over previous state-of-the-art. Gender-specific models improved emotion recognition by 5.7% UAR,… [145] (2022)Building and benchmarking the motivated deception corpus: Improving the quality of deceptive text through gaming
Behavior Research Methods State-of-the-art BERT classifiers that achieved 93.6% accuracy on traditional deception corpora struggled with the motivated deception corpus, achieving only 76% accuracy within statement triplets and… [146] (2022)In detecting deception using Decision Tree and SVM across different cues
Nigerian Journal of Technology The fusion of visual and audio features achieved 97.6% accuracy in deception detection, significantly outperforming either verbal-only or nonverbal-only approaches, with non-verbal features proving particularly… [147] (2022)Hierarchical Bayesian modeling for time series physiological data during the autonomic-based concealed information test.
The Proceedings of the Annual Convention of the Japanese Psychological Association Hierarchical Bayesian models applied to 167 participants' CIT data provided more accurate effect size estimates than conventional methods, confirmed CIT validity, and revealed heart rate… [148] (2022)Vgg-16 And Vgg-19 Architecture Models In Lie Detection Using Image Processing
2022 6th International Conference on Information Technology, Information Systems and Electrical Engineering (ICITISEE) The Adam + VGG-16 combination achieved the best performance among 18 tested configurations, demonstrating that the shallower VGG-16 architecture with adaptive moment estimation optimization provides… [149] (2022)EEG Guided Multimodal Lie Detection with Audio-Visual Cues
2022 2nd International Conference on Artificial Intelligence (ICAI) The multimodal fusion approach combining EEG, video-based facial movement analysis via optical flow, and attention-based acoustic processing demonstrated enhanced deception detection capability compared to single-modality… [150] (2022)A Personal Model of Trumpery: Linguistic Deception Detection in a Real-World High-Stakes Setting
Psychological Science The personalized model achieved 73% accuracy in training data and 74% in out-of-sample testing, with 36 linguistic variables showing significant differences between truthful and deceptive… [151] (2022)Lie Detection: Truth Identification from EEG Signal Using Frequency and Time Features with SVM Classifier
Wasit Journal of Computer and Mathematics Science The system achieved 83% overall accuracy, with specific channels (8 and 12) reaching 100% accuracy when using 16-block segmentation and statistical moment features with SVM… [152] (2022)Multimodal Deception Detection Using Real-Life Trial Data
IEEE Transactions on Affective Computing The multimodal system achieved 83.05-84.18% accuracy in detecting deception at the subject level, with the best performance obtained by combining visual and acoustic features, significantly… [153] (2022)The Application of Machine Learning in Deception Detection : a study based on the effect of empathy on deception detection
2022 4th International Conference on Machine Learning, Big Data and Business Intelligence (MLBDBI) The research investigates whether empathy levels moderate human performance when using ML-based deception detection systems, potentially informing personalized approaches to human-AI collaboration in credibility assessment. [154] (2022)Deception Detection Towards Multi-turn Question Answering with Context Selector Network
Lecture Notes in Computer Science The Context Selector Network outperformed state-of-the-art baseline models by leveraging contextual information across multi-turn conversations to automatically identify deceptive signals. Specific performance metrics are not… [155] (2022)Automatic deception detection system based on hybrid feature extraction techniques
Indonesian Journal of Electrical Engineering and Computer Science The hybrid deception detection system achieved 88.06% classification accuracy by combining FACS-based action units, head movement patterns, and eye gaze directional features from video recordings… [156] (2022)Attempt at Detection of Deception Based on Records of Physiological Reactions Remotely Captured with FaceReader Software: Part 1
European Polygraph The study represents a pioneering attempt to validate automated facial expression analysis for deception detection, though specific diagnostic performance results are not provided in this… [157] (2022)The emergence of “truth machines”?: Artificial intelligence approaches to lie detection
Ethics and Information Technology The paper's central argument is that HR managers should cease using technologically-based lie detection systems entirely and work to foster trust and accountability on a… [158] (2022)Dataset Bias in Deception Detection
2022 26th International Conference on Pattern Recognition (ICPR) A gender-prediction classifier achieved comparable performance (64.6% and 59.3% AUC) to state-of-the-art deception detection systems on the Real-life Trial dataset, while those same techniques performed… [159] (2022)Progress of Keystroke Dynamics in Deception Research
Fa Yi Xue Za Zhi (Journal of Forensic Medicine) Keystroke dynamics research has achieved 75-95% accuracy in detecting deception across various contexts, with studies using unexpected questions and machine learning classification reaching up to… [160] (2022)Machine Learning Based Deception Detection System in Online Social Networks
International Journal of Pure and Applied Sciences Seven supervised ML methods demonstrated varying performance in deception detection across two datasets when evaluated using four different metrics, providing empirical benchmarks for algorithm selection… [161] (2022)LDM: A Systematic Review on Lie detection Methodologies
The review identified a clear evolution from traditional contact-based polygraph methods to machine learning-driven contactless approaches utilizing facial micro-expressions, thermal imaging, and voice analysis, with… [162] (2022)Attention-Aware Multi-modal RNN for Deception Detection
2022 IEEE International Conference on Big Data (Big Data) The attention-aware multimodal RNN achieved 96% accuracy on the Real-Life Trial dataset, substantially exceeding human performance (50%) and earlier multimodal systems (83%), while providing interpretable… [163] (2022)Intentional-Deception Detection Based on Facial Muscle Movements in an Interactive Social Context
Pattern Recognition Letters Facial EMG achieved 90% accuracy in detecting intentional deception, with the corrugator supercilii and zygomaticus muscles playing key roles. Muscles related to lie detection were… [164] (2022)The politics of deceptive borders: 'biomarkers of deceit' and the case of iBorderCtrl
Information, Communication & Society The iBorderCtrl system showed severe limitations including training on only 32 actors with demographic bias, detection accuracy of only 75%, and methodologically invalid validation procedures… [165] (2022)Recurrent Neural Networks for Deception Detection in Videos
Communications in Computer and Information Science The Stacked GRU model achieved exceptional performance with 93.69% accuracy and an AUC of 0.9853, surpassing all other tested architectures and demonstrating that facial landmark… [166] (2022)Applying Neural Networks in Polygraph Testing
Journal of Computer and Systems Sciences International The voting ensemble neural network approach demonstrated increased efficiency in polygraph testing and decreased erroneous conclusions by combining multiple classifier predictions to analyze standard psychophysiological… [167] (2021)Man versus Machine: Comparing manual with LIWC coding of perceptual and contextual details for verbal lie detection
When set to the same metric, neither manual nor LIWC coding outperformed the other in discriminating truthful from deceptive statements (effect sizes ranged from d… [168] (2021)The Role of Lie Detection Based System in Controlling Borders
DAAAM Proceedings Laboratory testing of the iBorderCtrl ADDS component with 32 participants achieved approximately 76% accuracy (75.56% for deceptive, 73.67% for truthful participants), but field testing produced… [169] (2021)Explaining Neural Network Results by Sensitivity Analysis for Deception Detection
Communications in Computer and Information Science Neural networks analyzing observer physiological responses achieved 62% deception detection accuracy, exceeding typical human performance of 54%. Sensitivity analysis combined with genetic algorithms successfully extracted… [170] (2021)How humans impair automated deception detection performance
Acta Psychologica Machine learning alone achieved 69% accuracy in detecting deception, but human involvement through hybrid decision-making reduced performance to chance level (50%), demonstrating that human judgment… [171] (2021)Lie Detection Based on a DIFCW Radar with Machine Learning
2021 IEEE MTT-S International Wireless Symposium (IWS) Specific findings unavailable; paper proposes contactless radar-based deception detection using physiological signal analysis with machine learning. [172] (2021)Implementation of Muscle Testing for Lie Detection
Advances in Science, Technology & Innovation Traditional lie detection methods produce false positives when subjects are anxious or emotionally aroused, and computational approaches to lie prediction remain underdeveloped, creating opportunity for… [173] (2021)Use of Machine Learning for Deception Detection From Spectral and Cepstral Features of Speech Signals
IEEE Access The LSTM classification method with time-difference spectral energy features achieved the highest deception detection accuracy after PCA dimensionality reduction, outperforming traditional MFCC-based approaches and demonstrating… [174] (2021)Reconciling Adapted Psychological Profiling with the New European Data Protection Legislation
Studies in Computational Intelligence The paper identifies a fundamental challenge in reconciling black-box artificial neural network systems with GDPR's requirement that individuals receive meaningful explanations of automated profiling decisions,… [175] (2021)Affect-Aware Machine Learning Models for Deception Detection
Proceedings of the AAAI Conference on Artificial Intelligence Multimodal models incorporating facial affect achieved 91% AUC for deception detection, while unsupervised approaches using facial valence reached 80% AUC, outperforming human ability and validating… [176] (2021)Brain-based concealed memory detection is driven mainly by orientation to salient items
Cortex P3 CIT effects were similarly strong in both conceal and reveal conditions and enhanced for high-salience items, demonstrating that brain-based concealed memory detection is driven… [177] (2021)Automatic Deception Detection using Multiple Speech and Language Communicative Descriptors in Dialogs
APSIPA Transactions on Signal and Information Processing The multi-modal model achieved 80.61% accuracy detecting deception in Mandarin dialogs, with analysis revealing that deceiver behaviors can be observed through interrogator responses in conversational… [178] (2021)Deception Detection Using Machine Learning
Proceedings of the Annual Hawaii International Conference on System Sciences Multiple machine learning algorithms successfully detected deception in reviews, news articles, and emails, with SVM and Random Forest showing strong performance. The research identified shared… [179] (2021)Keynote Speech 4: A Conversational AI Approach to Detecting Deception and Tackling Insurance Fraud
2021 Tenth International Conference on Intelligent Computing and Information Systems (ICICIS) The underlying Decision Engine achieved 69-72% accuracy detecting deception in conversations using linguistic markers including hedging, pronoun usage, sentiment variation, and proximity-based features, outperforming baseline… [180] (2021)Lie to my face: An electromyography approach to the study of deceptive behavior
Brain and Behavior Machine learning analysis of facial muscle activity achieved 73% accuracy in detecting lies, revealing two distinct groups: those who betrayed lies through cheek muscle activation… [181] (2021)Affect-Aware Deep Belief Network Representations for Multimodal Unsupervised Deception Detection
2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021) The unsupervised model trained on facial valence and visual features achieved 80% AUC (70% accuracy, 88% precision), substantially exceeding typical human deception detection performance of… [182] (2021)Development of Non-contact Deception Detection using Facial Expression and Analysis of Significant Time Period
The Korean Academy of Scientific Criminal Investigation The facial expression-based system achieved 85.19% accuracy in classifying deceptive versus truthful videos on test data, confirming the viability of non-contact deception detection and identifying… [183] (2021)Development of Spectral Speech Features for Deception Detection Using Neural Networks
2021 IEEE 12th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON) Neural networks achieved 91.67% accuracy using Bark scale spectral energy features on police interrogation recordings, with delta and time-difference energy features proving critical for distinguishing… [184] (2021)Brain-Computer Interface in Lie Detection
Advances in Intelligent Systems and Computing The review demonstrates that P300-based BCI lie detection methods achieve accuracy rates of 74-86% for distinguishing guilty from innocent subjects, representing a neurophysiologically-grounded alternative to… [185] (2021)A Novel Hybrid Network Model Based on Attentional Multi-Feature Fusion for Deception Detection
IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences The attention-based multi-feature fusion approach significantly outperformed traditional fusion methods and achieved advanced performance using only small amounts of labeled speech data, demonstrating the viability… [186] (2021)A Look Into the Power of fNIRS Signals by Using the Welch Power Spectral Estimate for Deception Detection
Frontiers in Human Neuroscience Spontaneous deceptive behavior produced significantly higher power than controlled behavior in frontopolar area (BA10) and DLPFC (BA09), with statistical differences (p < 0.05 and p… [187] (2021)Prelude: Relational Communication and the Link to Deception
Terrorism, Security, and Computation Trust was the most powerful predictor of deception, with all three relational message dimensions independently distinguishing Spies from Villagers. Deceivers failed to show the increased… [188] (2021)Lie Detection by Facial Expressions in Real Time
2021 International Conference on Decision Aid Sciences and Application (DASA) The study demonstrates feasibility of real-time automated lie detection through facial expression analysis, identifying specific facial cues including eye blinking, lip movements, head movement, and… [189] (2021)Identity Unbiased Deception Detection by 2D-to-3D Face Reconstruction
2021 IEEE Winter Conference on Applications of Computer Vision (WACV) The identity-unbiased approach achieved 68% accuracy on high-stakes courtroom deception detection with 0.72 recall, demonstrating improved generalization by separating temporal behavioral cues from person-specific facial… [190] (2020)Phychoacoustic Masking of Delta and Time -Difference Cepstrum Features for Deception Detection
2020 11th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) Delta cepstrum and time-difference cepstrum features extracted using psychoacoustic masking principles and triangular critical band filters successfully distinguished truthful from deceptive speech in a real… [191] (2020)Novel Lie Speech Classification by using Voice Stress
Proceedings of the 12th International Conference on Agents and Artificial Intelligence The LSTM-based model achieved 72.5% precision in detecting deceptive speech from voice stress patterns, demonstrating that acoustic features alone can provide discriminative information for lie… [192] (2020)Moth Monarch Optimization-Based Deep Belief Network in Deception Detection System
Sādhanā The MMO-DBN system achieved exceptionally high classification performance with 98.4% accuracy, 98.36% sensitivity, and 93.75% specificity in deception detection. [193] (2020)Robust Deep Learning Ensemble against Deception
IEEE Transactions on Dependable and Secure Computing XEnsemble achieves high defense success rates against eleven adversarial attack methods and effectively detects out-of-distribution inputs, outperforming existing defense methods through a unified framework that… [194] (2020)Interview based connectivity analysis of EEG in order to detect deception
Medical Hypotheses Brain connectivity analysis during naturalistic interviews achieved 86.25% accuracy in detecting deception, with guilty subjects showing increased connectivity in the right hemisphere, especially the right… [195] (2020)Phase synchrony and its application to lie detection
2020 IEEE International Conference on Power, Intelligent Computing and Systems (ICPICS) Pattern recognition analysis of phase synchrony achieved 88% accuracy in lie detection, with deceptive responses showing greater oscillatory synchronization than truthful responses across brain regions… [196] (2020)Application of Machine Learning in Deception Detection
Advances in Intelligent Systems and Computing The neurofuzzy model achieved 97.1% accuracy with nonverbal cues alone, outperforming verbal-only (84.3%) and combined approaches (92.5%), while decision tree classifiers reached 97.6% accuracy on… [197] (2020)Effects of Pre-trained Word Embeddings on Text-based Deception Detection
2020 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech) Domain-specific word embeddings outperformed general-purpose pre-trained embeddings for deception detection, challenging the assumption that larger pre-trained models universally improve performance. The combination of embedding technique… [198] (2020)Deception in the eyes of deceiver: A computer vision and machine learning based automated deception detection
Expert Systems with Applications Eye movements emerged as the most significant predictors of deception, with Random Forest classification achieving 78% accuracy using the 24 most important behavioral features extracted… [199] (2020)Automated deception detection of males and females from non-verbal facial micro-gestures
IEEE World Congress on Computational Intelligence - IJCNN 2020 Both males and females showed reduced deception detection accuracy when classified by mixed-gender trained models rather than gender-specific classifiers, revealing distinct nonverbal behavioral patterns between… [200] (2020)An Audio-Visual based Feature Level Fusion Approach Applied to Deception Detection
Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications Feature-level fusion of audio and visual modalities improved deception detection accuracy compared to using either modality in isolation, with systems aggregating features from multiple layers… [201] (2020)Leveraging Shared and Divergent Facial Expression Behavior Between Genders in Deception Detection
2020 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020) GAHL outperformed logistic regression, decision trees, and SVM models in small-sample deception detection tasks (N<75), successfully balancing the use of full datasets for shared behavioral… [202] (2020)1, 2, 3 crimes you're out: Ocular‐motor methods for detecting deception in a multiple‐issue screening protocol
Journal of Forensic Sciences Classification accuracy ranged from 83.3% to 91.7% across participant groups, with the system also demonstrating strong item-level accuracy in identifying specific deceptive topics (81.1-88.3% accuracy). [203] (2020)Video Based Deception Detection Using Deep Recurrent Convolutional Neural Network
Communications in Computer and Information Science The deep recurrent convolutional neural network architecture outperformed four existing state-of-the-art techniques in detecting deception from real-life trial video footage by effectively capturing spatiotemporal behavioral… [204] (2020)Airport artificial intelligence can detect deception: or am i lying?
Security Journal The authors conclude that in vivo testing of AI deception detection based on micro-expressions is naïve and misinformed, noting that micro-expressions occur in only 2%… [205] (2020)Classification of EEG signals using hybrid combination of features for lie detection
Neural Computing and Applications The hybrid combination of EMD-derived features with Support Vector Machine classification achieved 99.44% accuracy in detecting deception, significantly outperforming traditional single-domain feature extraction methods. [206] (2020)Understanding and Explicitly Measuring Linguistic and Stylistic Properties of Deception via Generation and Translation
Proceedings of the 13th International Conference on Natural Language Generation When conditioned on topic, deceptive content is shorter, less readable, more biased, and more subjective than credible content. Style transfer from deceptive to credible writing… [207] (2020)Technologies in the twilight zone: early lie detectors, machine learning and reformist legal realism
International Review of Law, Computers & Technology The study reveals a dual-system phenomenon where polygraphs were consistently excluded from courtrooms under the Frye standard yet flourished in employment screening, investigation, and interrogation—a… [208] (2020)Truth Identification from EEG Signal by using Convolution neural network: Lie Detection
2020 43rd International Conference on Telecommunications and Signal Processing (TSP) The CNN approach achieved 84.44% overall accuracy in EEG-based truth/lie classification, with specific accuracies of 91% for detecting truthful statements and 89% for identifying deceptive… [209] (2020)Multimodal Deception Detection Using Automatically Extracted Acoustic, Visual, and Lexical Features
Interspeech 2020 The automated multimodal system achieved 73% accuracy on conversational deception detection, significantly outperforming previous manual annotation methods and human performance on the same dataset. [210] (2020)Identifying the optimal features in multimodal deception detection
Multimodal Technologies and Interaction Thermal imaging achieved 90.1% accuracy in mock crime scenarios compared to 70.9% for traditional GSR/PPG measures. Perinasal and cheek regions showed superior discriminatory power, and… [211] (2020)Gender-based multimodal deception detection
Multimodal Technologies and Interaction Classification accuracies of 90.1% for mock crime and 74.7% for best friend scenarios were achieved, with perinasal and cheek areas showing the greatest discriminatory power… [212] (2019)On Human Predictions with Explanations and Predictions of Machine Learning Models: A Case Study on Deception Detection
Proceedings of the Conference on Fairness, Accountability, and Transparency Explanations alone produced only modest improvements in human deception detection (from 51.1% baseline), while showing ML predicted labels yielded over 20% relative improvement. However, this… [213] (2019)Detecting Deception in Political Debates Using Acoustic and Textual Features
2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) BERT textual features achieved MAE of 0.76, while i-vector acoustic features achieved comparable performance with MAE of 0.77. Multimodal fusion combining acoustic and textual features… [214] (2019)P300-based deception detection of mock network fraud with modified genetic algorithm and combined classification
2019 7th International Winter Conference on Brain-Computer Interface (BCI) The modified genetic algorithm and combined classification approach effectively detected concealed information about network fraud activities, demonstrating the feasibility of applying P300-based testing to cybercrime… [215] (2019)Analysis of Concealed Anger Emotion in a Neutral Speech Signal
2019 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics (DISCOVER) Statistical measures of autocorrelated pitch, including central tendency and measures of spread, can detect and characterize concealed anger emotion in synthesized speech signals that combine… [216] (2019)Region of Interest Based Comparative Analysis in Deception Detection using Thermal Images
2019 2nd International Conference on Intelligent Computing, Instrumentation and Control Technologies (ICICICT) The study identified which specific facial regions provide superior discriminatory power for deception detection through comparative thermal analysis, contributing to standardization efforts in thermal imaging-based… [217] (2019)Mapping of the brain activation associated with deception using fused EEG and fNIRS
Neural Imaging and Sensing 2019 The combined EEG-fNIRS approach achieved 0.94 AUC for distinguishing guilty from innocent participants, significantly outperforming P300 alone (0.85 AUC) or fNIRS alone (0.84 AUC). Guilty… [218] (2019)Detecting Lie-A Practical Approach
Revue d'Intelligence Artificielle The study successfully demonstrated that physiological signals can be measured using accessible biomedical equipment to create a functional personal polygraph, identifying gender-based statistical differences in… [219] (2019)Real Time Deception Detection for Criminal Investigation
2019 National Information Technology Conference (NITC) The PREDICTOR system demonstrated feasibility for real-time monitoring of deception, emotion, and attentiveness during criminal investigations, though specific accuracy metrics are not available in accessible… [220] (2019)Face-Focused Cross-Stream Network for Deception Detection in Videos
2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) The FFCSN achieved state-of-the-art performance on real-life courtroom trial videos (104 videos, 58 subjects) by explicitly detecting faces in the spatial stream and performing cross-stream… [221] (2019)A Synergistic Concealed Information Test With Novel Approach for EEG Channel Selection and SVM Parameter Optimization
IEEE Transactions on Information Forensics and Security The dual Bat Algorithm approach successfully optimized both EEG channel selection and SVM hyperparameters simultaneously, achieving high classification accuracy while identifying that occipital lobe channels… [222] (2019)Speech Deception Detection Algorithm Based on SVM and Acoustic Features
2019 IEEE 7th International Conference on Computer Science and Network Technology (ICCSNT) The AF-SVM algorithm successfully demonstrated automated deception detection from speech using Support Vector Machine classification of acoustic features, validating the feasibility of machine learning approaches… [223] (2019)Detecting Deception Using Neurofuzzy Approach
International Journal of Intelligent Computing Research The neurofuzzy model achieved 97.1% accuracy detecting deception from nonverbal features alone, significantly outperforming both human judges and chance-level detection. Different algorithms performed optimally with… [224] (2019)Automatic Long-Term Deception Detection in Group Interaction Videos
2019 IEEE International Conference on Multimedia and Expo (ICME) The system achieved AUCs over 0.70 for identifying spies in extended group interactions. Low-level and high-level video features alone were insufficient; combining them with LiarRank… [225] (2019)Bag-of-Lies: A Multimodal Dataset for Deception Detection
IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops Multimodal fusion significantly outperformed single-modality approaches, with the best accuracy of 70% achieved by combining video, audio, and EEG data, while gaze data performed best… [226] (2019)Box of Lies: Multimodal deception detection in dialogues
Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies Multimodal machine learning classification achieved 69% accuracy in detecting conversational deception, exceeding both random baseline (50%) and human performance, with dialogue features providing unique predictive… [227] (2019)The intra-individual consistency of the differential physiological responses on the Concealed Information Test: hierarchical Bayesian approach.
The Proceedings of the Annual Convention of the Japanese Psychological Association Hierarchical Bayesian estimates of effect sizes were more accurate than conventional approaches, revealing that traditional Z-score methods overestimate the magnitude of differential responding. Heart rate… [228] (2019)Improved semi-supervised autoencoder for deception detection
PLOS ONE The SS-ANE model achieved 61.81% and 63.89% accuracy on the Killer corpus with only 100 and 200 labeled examples respectively, outperforming baseline models even when… [229] (2019)Comparison of different functional connectives based on EEG during concealed information test
Biomedical Signal Processing and Control Deception responses showed significantly increased functional connectivity and enhanced small-world network properties compared to truth-telling, with mutual information-based connectivity analysis outperforming other linear and nonlinear… [230] (2019)Detecting Concealed Information in Text and Speech
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics Multi-task learning framework with acoustic, linguistic, and individual features outperformed human detection of concealed information by over 15%. Acoustic-prosodic indicators including pitch, intensity, and speaking… [231] (2019)Robust Algorithm for Multimodal Deception Detection
2019 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR) The multimodal system combining audio (cepstral coefficients with SRKDA), text (bag-of-n-grams with SVM), and non-verbal features (AdaBoost classifier) achieved 97% deception detection accuracy through majority… [232] (2019)A Study on the Lie Detection of Telephone Voices Using Support Vector Machine
Studies in Computational Intelligence The researchers successfully demonstrated the application of SVM machine learning to classify telephone voices as truthful or deceptive based on formant and voice reliability parameters,… [233] (2019)Automatic deception detection in RGB videos using facial action units
Proceedings of the 13th International Conference on Distributed Smart Cameras (ICDSC) The automated facial analysis system achieved 76.84% accuracy in deception detection using only facial Action Units, demonstrating the potential of non-invasive video-based deception detection. [234] (2019)Event Related Potential (ERP) based Lie Detection using a Wearable EEG headset
2019 16th International Bhurban Conference on Applied Sciences and Technology (IBCAST) The study achieved 78.57% classification accuracy for detecting concealed information using a single-channel wearable EEG device, demonstrating feasibility of portable, consumer-grade EEG technology for deception… [235] (2018)Research of Brain Network on Deception Identification Using Phase Synchrony
2018 5th International Conference on Information Science and Control Engineering (ICISCE) The phase synchrony approach achieved 88% classification accuracy in distinguishing deceptive from truthful responses, demonstrating that functional connectivity network analysis can reveal meaningful brain communication… [236] (2018)Using Blink Rate to Detect Deception: A Study to Validate an Automatic Blink Detector and a New Dataset of Videos from Liars and Truth-Tellers
Journal of Imaging The automated system achieved 96.15% accuracy in classifying deceptive versus truthful responses based on blink rate patterns, with 99.3% accuracy in detecting individual blinks on… [237] (2018)What mechanisms underlie ERP-based concealed information detection?
International Journal of Psychophysiology P3 responses and multivariate brain patterns showed equivalent strength whether participants concealed or revealed information, but were strongly influenced by stimulus salience, demonstrating that brain-based… [238] (2018)Constructing the lie detection system with fuzzy reasoning approach
Granular Computing The fuzzy reasoning approach achieved 89.5% lie detection accuracy and successfully identified specific EEG frequency bands associated with deception, comparing favorably to other data mining… [239] (2018)Covert lie detection using keyboard dynamics
Scientific Reports Keystroke analysis distinguished liars from truth-tellers with 95% accuracy, with liars showing significantly longer reaction times, more typing errors, and slower typing speeds, especially when… [240] (2018)Do You Know Him? Gaze Dynamics Toward Familiar Faces on a Concealed Information Test
Journal of Applied Research in Memory and Cognition The study achieved 89% accuracy in detecting concealed face recognition using eye-tracking. Participants showed an orientation-avoidance pattern (initial fixation on familiar faces followed by avoidance)… [241] (2018)Multivariate data analysis of polygraph charts for assessment of biophysical characteristics
Izv. AltGU Both PCA and PLS methods successfully analyzed polygraph data to assess human biophysical characteristics and identify physiological response anomalies, with potential applications in insider threat… [242] (2018)Detecting Concealed Information with Fused Electroencephalography and Functional Near-infrared Spectroscopy
Neuroscience Combined EEG-fNIRS achieved AUC of 0.94 for differentiating guilty from innocent participants, compared to 0.85 for P300 alone and 0.84 for fNIRS alone, demonstrating superior… [243] (2018)Intelligent Deception Detection through Machine Based Interviewing
IEEE International Joint Conference on Neural Networks (IJCNN) Neural network classifiers successfully distinguished truthful from deceptive behavior in avatar interview scenarios, demonstrating that machine-based interviewing can elicit analyzable behavioral data for credibility assessment. [244] (2018)Deception Detection and Analysis in Spoken Dialogues based on FastText
2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) The FastText-based deception detector achieved significantly higher accuracy than human judges in distinguishing truthful from deceptive dialogue utterances. Combining acoustic features with the text-based model… [245] (2018)Automated Verbal Credibility Assessment of Intentions: The Model Statement Technique and Predictive Modeling
Applied Cognitive Psychology Machine learning classification achieved 80% accuracy in initial testing but dropped to 63% in independent validation, with liars paradoxically including more person and location references… [246] (2018)Deep Personality Recognition for Deception Detection
Interspeech 2018 Multi-task learning models that incorporated personality recognition improved deception detection F1 scores from 0.68 to 0.744, representing a 6%+ relative improvement and new state-of-the-art results… [247] (2018)Multimodal deception detection
The Handbook of Multimodal-Multisensor Interfaces: Signal Processing, Architectures, and Detection of Emotion and Cognition Multimodal fusion of linguistic, physiological, and thermal features achieved detection accuracies of 60-75%, with thermal forehead features outperforming other facial features. [248] (2018)Comparative Analysis of Classification Methods for Automatic Deception Detection in Speech
Lecture Notes in Computer Science Random Forest achieved the highest performance with 79.3% UAR, while k-Nearest Neighbor combined with ensemble methods reached 76.3% UAR, substantially exceeding chance-level performance. [249] (2018)Speech Based Deception Detection Using Bispectral Analysis
Proceedings of the 2nd International Conference on Graphics and Signal Processing Deceptive speech was accurately detected using normalized bispectral entropy for envelope (ENV) data and maximum bispectrum for temporal fine structure (TFS) data, with the authors… [250] (2018)Embodied Conversational Agent-Based Deception Detection
Lecture Notes in Networks and Systems Electrodermal activity successfully confirmed the presence of arousal during agent-based interviews, demonstrating that automated avatar systems can elicit valid physiological responses for detecting guilty knowledge… [251] (2017)Deception detection of EEG-P300 component classified by SVM method
Proceedings of the 6th International Conference on Software and Computer Applications The SVM classifier achieved 70.83% accuracy in distinguishing deceptive from truthful responses with extremely fast processing time (0.0283 seconds). Deceptive responses showed larger P300 amplitude… [252] (2017)Lie detection on pupil size by back propagation neural network
Procedia Computer Science The neural network with 10 hidden neurons achieved 100% accuracy in classifying pupil images during training, though results were not validated on independent test data… [253] (2017)Cross-domain deception detection using support vector networks
Soft Computing Within-domain accuracy reached 86%, but cross-domain performance dropped dramatically to 52-64%, demonstrating that linguistic deception cues do not transfer reliably across different topics and genres. [254] (2017)Social Signals of Deception and Dishonesty
Social Signal Processing Automated systems analyzing constellations of nonverbal behavioral cues can classify deception and truth more accurately than many human observers, with eye movements, kinesic patterns, and… [255] (2017)Deception Detection and Opinion Spam
Socio-Affective Computing The review demonstrates that affective and psychometric linguistic features effectively distinguish deceptive from truthful communication, with character n-gram models and low-dimensionality representations achieving competitive accuracy… [256] (2017)Using named entities for computer-automated verbal deception detection
Journal of Forensic Sciences Named entities discriminated truthful from deceptive hotel reviews with AUC values of 0.59-0.63, outperforming lexicon-based word counting (LIWC) and sentence specificity measures. Dates, money references,… [257] (2017)The application of microexpression in the detection of deception
Advances in Psychological Science Even when liars deliberately attempt to control their expressions, they cannot prevent microexpression leakage of true emotions, making microexpressions a theoretically valid tool for detecting… [258] (2017)Deception detection in Russian texts
Proceedings of the Student Research Workshop at the 15th Conference of the European Chapter of the Association for Computational Linguistics The classifier achieved 68.3% accuracy in distinguishing truthful from deceptive Russian texts, with statistically significant differences found across linguistic parameters. Model accuracy was found to… [259] (2017)Deep learning driven multimodal fusion for automated deception detection
2017 IEEE Symposium Series on Computational Intelligence (SSCI) The deep learning multimodal fusion approach achieved 96% prediction accuracy compared to 82% in previous literature, with early fusion outperforming late fusion by approximately 5%. [260] (2017)A Method to Extract P300 EEG Signal Feature Using Independent Component Analysis (ICA) for Lie Detection
Journal of Energy, Mechanical, Material, and Manufacturing Engineering ICA successfully extracted P300 components from multi-channel EEG recordings, with responses clustering in the 13-30 Hz beta frequency range associated with active cognitive processing. Amplitude… [261] (2017)Detecting deceptive behavior via integration of discriminative features from multiple modalities
IEEE Transactions on Information Forensics and Security The combination of thermal and linguistic modalities exceeded 70% accuracy in detecting deception, with forehead thermal features and combined unigram-LIWC linguistic features providing the strongest… [262] (2017)A brain signal-based credibility assessment approach
2017 Joint 17th World Congress of International Fuzzy Systems Association and 9th International Conference on Soft Computing and Intelligent Systems (IFSA-SCIS) The study demonstrated an alternative to P300-based deception detection using CSP feature extraction combined with logistic regression classification, though specific accuracy rates were not reported… [263] (2016)ID-Check: Online Concealed Information Test Reveals True Identity
Journal of Forensic Sciences The online ID-Check achieved near-perfect diagnostic accuracy with AUC = 0.98 and overall accuracy ranging from 86-94% in detecting individuals attempting to conceal their true… [264] (2016)Decoding fMRI signatures of real-world autobiographical memory retrieval
Journal of Cognitive Neuroscience Whole-brain multivoxel pattern analysis achieved near-perfect accuracy distinguishing recognized from novel events at 1-3 week retention intervals, successfully differentiated recollection from familiarity, and showed that… [265] (2016)Estimating the amenibility of new domains for deception detection
Proceedings of the Second Workshop on Computational Approaches to Deception Detection The study showed that out-of-rank distance measures on n-grams can predict whether new text domains will support successful deception classification, offering researchers a preliminary assessment… [266] (2016)Is Deception Emotional? An Emotion-Driven Predictive Approach
Interspeech 2016 Arousal emerged as the strongest individual emotion dimension for identifying deception, and emotion-related features derived from transfer learning successfully improved upon baseline acoustic features for… [267] (2016)Special Issue: Information Systems for Deception Detection
Journal of Management Information Systems The featured research demonstrates that computational approaches combining verbal, nonverbal, behavioral, and physiological cues significantly outperform human judgment in detecting deception across multiple domains, from… [268] (2016)Comparing a Perceptual and an Automated Vision-Based Method for Lie Detection in Younger Children
Frontiers in Psychology Judges detected lies above chance level but with modest accuracy. Increased movement was correlated with perception of lies, though automated analysis found no actual movement… [269] (2016)Prosodic Cues and Answer Type Detection for the Deception Sub-Challenge
Interspeech 2016 The system achieved UAR = 74.9% (winning performance), representing 13.0% improvement on development set and 3.8% on test set compared to the 68.3% baseline, by… [270] (2016)Using the verifiability of details as a test of deception: A conceptual framework for the automation of the verifiability approach
Proceedings of the Second Workshop on Computational Approaches to Deception Detection The paper provides theoretical and methodological blueprints for automating verifiability coding through LIWC-based linguistic analysis and rule-based named entity recognition, establishing a foundation for computational… [271] (2016)Deception Detection and Rumor Debunking for Social Media
The SAGE Handbook of Social Media Research Methods Human deception detection achieves only 50-63% accuracy without technological assistance. Automated machine learning approaches reach approximately 65% accuracy, with hybrid methods combining linguistic analysis and… [272] (2016)Exploring time- and frequency-dependent functional connectivity and brain networks during deception with single-trial event-related potentials
Scientific Reports Deceptive responses showed significantly greater connectivity than truthful responses, particularly in the theta frequency band between prefrontal/frontal and central regions and between prefrontal/frontal and left… [273] (2016)Untangling a Web of Lies: Exploring Automated Detection of Deception in Computer-Mediated Communication
Journal of Management Information Systems The automated detection algorithm achieved 70% accuracy in identifying deception, substantially outperforming the 54% accuracy of human judges. Deceivers minimized referencing and self-deprecation, included more… [274] (2015)A corpus-based analysis of potential linguistic indicators of corporate deception in tobacco industry documents
Cognitive-emotional verbs (believe, think, seem, feel) were strongly associated with likely deception and most prevalent in external audience documents and materials from the Tobacco Institute… [275] (2015)Deception is in the eye of the communicator: Investigating pupil diameter variations in automated deception detection interviews
IEEE International Conference on Intelligence and Security Informatics Pupil diameter changes over time were indicative of deception regardless of whether incriminating target items were shown, with deceivers showing distinct growth trajectories that differentiated… [276] (2015)An experiment of lie detection based EEG-P300 classified by SVM algorithm
2015 International Conference on Automation, Cognitive Science, Optics, Micro Electro-Mechanical System, and Information Technology (ICACOMIT) The SVM classifier achieved 70.83% accuracy in distinguishing deceptive from truthful responses with a computation time of 0.0283 seconds, demonstrating proof-of-concept for automated P300-based lie… [277] (2015)Human Deception Detection from Whole Body Motion Analysis
Whole-body motion capture analysis enables objective, quantitative assessment of deception-related behavior patterns across the entire kinematic profile, moving beyond subjective judgments and isolated behavioral observations… [278] (2015)Unobtrusive Deception Detection
The Oxford Handbook of Affective Computing Multimodal sensor fusion combining laser Doppler vibrometry with human judgment achieved 90% true positive rate with minimal false positives. An embodied conversational agent (AVATAR) system… [279] (2015)TESTING THE HYPOTHESIS OF CREATING A DIGITAL POLYGRAPH BASED ON VIDEO AND AUDIO DATA
Vestnik Gosudarstvennogo universiteta morskogo i rechnogo flota imeni admirala S. O. Makarova Specific findings are not available as the original full text could not be accessed. The research explored the feasibility of multimodal video-audio deception detection as… [280] (2015)Monte Carlo Estimates of the Validity of Polygraph Testing
Various Publications Used Monte Carlo simulation methods to estimate polygraph validity across different conditions. Statistical modeling confirmed high accuracy estimates consistent with empirical findings from field and… [281] (2015)A comparison of features for automatic deception detection in synchronous computer-mediated communication
2015 IEEE International Conference on Intelligence and Security Informatics (ISI) Structural features were independently effective in detecting deception, and combining structural features with linguistic features significantly improved deception detection performance compared to using linguistic features… [282] (2015)Are computers effective lie detectors? A meta-analysis of linguistic cues to deception
Personality and Social Psychology Review Relative to truth-tellers, liars showed greater cognitive load, expressed more negative emotions, distanced themselves more from events, used fewer sensory-perceptual words, and referred less to… [283] (2015)Trimodal Analysis of Deceptive Behavior
Proceedings of the 2015 ACM on Workshop on Multimodal Deception Detection Forehead thermal features provided the best single-region performance for deception detection. Multimodal fusion of thermal and linguistic features exceeded 70% accuracy, outperforming individual modalities and… [284] (2015)Distinguishing deception from non-deception in Chinese speech
2015 Sixth International Conference on Intelligent Control and Information Processing (ICICIP) Using acoustic features (pitch, formant, short-time energy, duration) and Gradient Boosting Decision Tree algorithms, researchers achieved 82-85% accuracy in detecting deception in Chinese speech, with… [285] (2015)Deception Detection using Real-life Trial Data
Proceedings of the 2015 ACM on International Conference on Multimodal Interaction Multimodal fusion of linguistic and gesture features achieved 60-75% accuracy in detecting deception from real courtroom footage, outperforming human judgment which showed only slight inter-rater… [286] (2015)Verbal and Nonverbal Clues for Real-life Deception Detection
Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing Multimodal machine learning models achieved 77-82% accuracy on real-world deception detection, significantly exceeding human performance (near chance-level), with SVM classifiers reaching 82% when combining linguistic… [287] (2015)Lie Detection from Speech Analysis Based on K–SVD Deep Belief Network Model
Lecture Notes in Computer Science The K-SVD deep belief network approach achieved 69.83% accuracy in deception detection and demonstrated that deep features extracted through sparse representation outperformed conventional acoustic features… [288] (2015)Single-trial Lie Detection Using a Combined fNIRS-Polygraph System
Frontiers in Psychology The combined fNIRS-polygraph system achieved 86.5% classification accuracy, significantly outperforming fNIRS alone (71.6%) or polygraph alone (74.5%), demonstrating that multimodal measurement captures complementary deception-related processes… [289] (2015)Detection of Deception in the Mafia Party Game
Proceedings of the 2015 ACM on International Conference on Multimodal Interaction The automated system achieved 0.61 AUC (70.26% accuracy at optimal duration) classifying deceptive versus truthful players based on facial action units, with eye features outperforming… [290] (2015)Neuro signal based lie detection
2015 IEEE International Symposium on Robotics and Intelligent Sensors (IRIS) The study achieved approximately 90% average accuracy in classifying lie versus truth responses, with individual participant accuracy ranging from 46% to 99% across testing rounds,… [291] (2015)Combination of event related potentials and peripheral signals in order to improve the accuracy of the lie detection systems
JSDP The combination of event-related potentials and peripheral cardiovascular signals improved lie detection system accuracy compared to single-modality approaches, validating multimodal physiological measurement in credibility assessment. [292] (2014)Exploratory Study of Relationships among Statement Credibility, Context, and Semantic Similarity
2014 IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and Intelligent Agent Technologies (IAT) The study demonstrated that semantic similarity measures could be systematically applied to sentence-level credibility assessment using a probabilistic framework, providing preliminary evidence that semantically related… [293] (2014)Deception Detection Within and Across Cultures
Theory and Applications of Natural Language Processing Cross-cultural deception classifiers achieved 60-70% accuracy using LIWC psycholinguistic features as a semantic bridge, with linguistic categories showing 2-4% improvement over unigrams. Universal deception cues… [294] (2014)Polygraph Test Results Assessment by Regression Analysis Methods
Science and Education of the Bauman MSTU The study demonstrated that regression-based statistical methods could provide systematic evaluation of polygraph data, reducing examiner subjectivity and establishing quantifiable decision rules for deception detection… [295] (2014)Facial Behavioral Analysis: A Case Study in Deception Detection
British Journal of Applied Science & Technology The human examiner achieved 56.25% accuracy while Logistic Regression achieved 47.9% sensitivity and 71.2% specificity, with machine learning showing bias toward predicting truthfulness. [296] (2014)Reaction Time
Encyclopedia of Deception Meta-analysis found lying produces a large RT delay (d = 1.049) compared to truth-telling, even after correcting for publication bias. The RT-CIT shows low false… [297] (2014)Keystroke Patterns as Prosody in Digital Writings: A Case Study with Deceptive Reviews and Essays
Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) Combined keystroke features with stylometry achieved over 90% classification accuracy, with deceptive writers showing increased pauses and revisions, particularly spending more time on first-person pronouns,… [298] (2014)Measurement Model for Deception Detection in Online Chat Software
Information Technology Journal The software analysis produced an average deception probability of 11.36% across participants, and users' trust in the analysis affected their decision-making during chat interactions. [299] (2014)Content analysis algorithms: an innovative and accurate approach to statement veracity assessment
European Polygraph MASAM with decision algorithms achieved 96.87% accuracy for truthful statements and 69.23% for false statements, dramatically outperforming traditional SVA, RM, and SCAN techniques which performed… [300] (2014)Patterns of nonverbal behavior associated with truth and deception: illustrations from three experiments
Journal of Nonverbal Behavior Deceptive and truthful participants exhibited significantly different temporal patterns of nonverbal behaviors across all three experiments, with patterns occurring far more frequently than chance. Deceivers… [301] (2014)A Rigidity Detection System for Automated Credibility Assessment
Journal of Management Information Systems The system successfully detected kinesic rigidity in real-time during concealed information tests, confirming rigidity as a viable automated deception indicator and providing new theoretical insights… [302] (2014)A Novel Algorithm to Enhance P300 in Single Trials: Application to Lie Detection Using F-Score and SVM
PLoS ONE The F-score_SVM model combining spatial denoising with optimized feature selection achieved over 96% accuracy in classifying guilty versus innocent subjects based on P300 responses to… [303] (2014)Cross-cultural Deception Detection
Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) Deception detection classifiers showed reduced but viable performance in cross-cultural conditions, with certain psycholinguistic word classes appearing selectively across cultures while some deceptive patterns remained… [304] (2014)Detection of Deception Via Handwriting Behaviors Using a Computerized Tool: Toward an Evaluation of Malingering
Cognitive Computation Deceptive writing was significantly broader and took longer than truthful writing. Three distinct handwriting profiles emerged with profile-specific spatial and temporal differences between deceptive and… [305] (2014)A novel classification method based on ICA and ELM: a case study in lie detection
Bio-Medical Materials and Engineering The ICA_ELM method achieved 95.40% classification accuracy with dramatically reduced computational time compared to backpropagation neural networks and support vector machines, demonstrating both high accuracy… [306] (2014)“Time for Some Traffic Problems”: Enhancing E-Discovery and Big Data Processing Tools with Linguistic Methods for Deception Detection
Journal of Digital Forensics, Security and Law The linguistic classifier successfully demonstrated capability to stratify the large Enron Email Dataset and identify potentially deceptive records by analyzing linguistic feature distributions while avoiding… [307] (2014)A Multimodal Dataset for Deception Detection
Proceedings of the Ninth International Conference on Language Resources and Evaluation Multimodal fusion combining video, audio, and EEG achieved 70% average accuracy using Random Forest classifiers, substantially outperforming individual modality performance and baseline polygraph accuracy rates… [308] (2014)Discourse cues to deception in the case of multiple receivers
Information & Management Statistical analysis validated that discourse-level features effectively distinguish deceivers from truth-tellers in multi-receiver contexts, identifying new deception behaviors such as disturbance and discourse features like… [309] (2014)Deception detection using a multimodal approach
Proceedings of the 16th International Conference on Multimodal Interaction The combination of thermal and linguistic features exceeded 70% detection accuracy for both deceptive and truthful classes, significantly outperforming single-modality approaches and demonstrating the viability… [310] (2014 • THERMAL ANALYSIS)Thermal Facial Analysis for Deception Detection
IEEE Journals Thermal imaging analysis of 492 responses from 25 participants. Achieved 87% prediction accuracy using within-person methodology, but between-person models did not generalize well. [311] (2014)Cues to Deception in Social Media Communications
2014 47th Hawaii International Conference on System Sciences Machine learning classifiers achieved 90% average accuracy in detecting deception using linguistic features including sentence complexity, sentiment, emoticon usage, and informality markers, with mean precision… [312] (2013)A Novel Approach for Lie Detection Based on F-Score and Extreme Learning Machine
PLoS ONE The F-score_ELM method achieved exceptional testing accuracies of 99.27% sensitivity and 98.17% specificity in classifying guilty versus innocent subjects, outperforming traditional PCA-based approaches while requiring… [313] (2013)Seeing through Deception: A Computational Approach to Deceit Detection in Spanish Written Communication
Linguistic Evidence in Security, Law and Intelligence The LIWC-based classifier achieved 84.5% F-measure, with linguistic and psychological process dimensions proving most discriminatory. LIWC psycholinguistic features outperformed Bag-of-Words models for detecting deception in… [314] (2013)Lie detection in interrogations using digital signal processing of brain waves
2013 3rd International Conference on Instrumentation, Communications, Information Technology and Biomedical Engineering (ICICI-BME) Initial results showed good reliability in the hardware-implemented DSP-based lie detection system, though the researchers acknowledged that further testing on additional subjects was required to… [315] (2013)Deception detection in speech using bark band and perceptually significant energy features
2013 IEEE 56th International Midwest Symposium on Circuits and Systems (MWSCAS) Bark scale spectral energy features outperformed traditional mel-frequency cepstral coefficients (MFCCs) for deceptive speech classification, demonstrating the viability of psychoacoustically-motivated features for voice-based deception detection. [316] (2013)Towards the Detection of Deception in Interactive Multimedia Environments
Lecture Notes in Computer Science The study successfully demonstrated deception detection using non-intrusive off-the-person sensors in an interactive game-based environment, providing useful guidelines for future work while engaging public audiences… [317] (2013)Extending Textual Models of Deception to Interrogation Settings
Linguistic Evidence in Security, Law and Intelligence The surprising finding was that removing question-language influence from responses degraded rather than improved deception detection, as deceptive individuals showed qualitatively different linguistic response patterns… [318] (2013)High Precision Credibility Analysis of Information on Twitter
2013 International Conference on Signal-Image Technology & Internet-Based Systems Testing on 2,960 tweets achieved kappa statistics greater than 0.6 agreement with human evaluators. The expertise score successfully addressed accuracy problems when tweet volume was… [319] (2013)The autobiographical IAT: a review
Frontiers in Psychology Validation experiments demonstrated over 90% accuracy in detecting true autobiographical memories, with test-retest agreement exceeding 90% and split-half reliability of 88%. The technique successfully detected… [320] (2013)A functional analysis of deception detection of a mock crime using infrared thermal imaging and the Concealed Information Test
Frontiers in Human Neuroscience Functional discriminant analysis achieved 98.89% classification accuracy. Guilty participants showed significantly higher periorbital temperatures when responding to crime-relevant vs. irrelevant questions, while innocent participants showed… [321] (2012)A game theory approach to deception strategy in computer mediated communication
2012 IEEE International Conference on Intelligence and Security Informatics Deceivers employ systematically different communication strategies than truth-tellers and adapt their behaviors across multiple interaction stages, supporting a rational-actor model of strategic deception that challenges… [322] (2012)An experimental comparison of real and artificial deception using a deception generation model
Decision Support Systems The Application Deception Model successfully generated artificial deception with characteristics similar to real deception, providing evidence that artificially generated data can reduce training data collection… [323] (2012)A Framework for Modeling Decision Making and Deception with Semantic Information
2012 IEEE Symposium on Security and Privacy Workshops The study proves that in the proposed three-player game, there are always equilibria in pure strategies, and in more complex scenarios involving uncertainty about statement… [324] (2012)Credibility assessment and inference for fusion of hard and soft information
Advances in Design for Cross-Cultural Activities Part I The research introduces an analytical framework enabling automatic and efficient processing of both physics-based sensor data and human-generated information for enhanced decision-making, with applications including… [325] (2012)Unification Of Randomized Anomaly In Deception Detection Using Fuzzy Logic Under Uncertainty
International Journal of Research in Computer Science Fuzzy logic outperformed randomness by approximately 45% and uncertainty methods by 55% in maximum deception detection scenarios, demonstrating that graded membership functions capture the ambiguous… [326] (2012)Algorithms for detecting concealed knowledge among groups when the critical information is unavailable
Journal of Experimental Psychology: Applied Reanalysis of three datasets showed that both averaging and PCA-based algorithms successfully detected critical items and differentiated knowledgeable from unknowledgeable subjects with efficiency approaching standard… [327] (2012)Deception detection for the tangled web
ACM SIGCAS Computers and Society The review identified substantial inconsistencies in deception cue measurements, contradictions across studies, and significant context-dependency in detection effectiveness, revealing that the field lacked sufficient consistency… [328] (2012)The design and development of a lie detection system using facial micro-expressions
2nd International Conference on Advances in Computational Tools for Engineering Applications (ACTEA) The system successfully classified eight basic facial expressions using geometric-based dynamic templates and frame-by-frame analysis, demonstrating technical feasibility of automated micro-expression detection, though specific accuracy… [329] (2012)A New Approach for Concealed Information Identification Based on ERP Assessment
Journal of Medical Systems Wavelet coefficient features showed statistically significant differences between concealed and irrelevant information (p < 0.05), with KPCA successfully reducing feature dimensionality while improving SVM classifier… [330] (2012)Linguistic Cues to Deception Assessed by Computer Programs: A Meta-Analysis
PsycEXTRA Dataset Computer-analyzed linguistic cues reliably differentiate lies from truths, with liars showing increased cognitive load, more negative emotion, greater psychological distancing, fewer sensory details, and reduced… [331] (2012)A Secure Group Collaboration Protocol for Nonverbal Human Social Signals Featuring Deception Detection
2012 International Conference on Privacy, Security, Risk and Trust and 2012 International Confernece on Social Computing The authors developed a cryptographic protocol allowing secure multi-party computation on deception-related nonverbal social signals, providing privacy guarantees against semi-honest participants while enabling collaborative analysis… [332] (2012)A New Lie Detection Method Based on Small-number of P300 Responses
2012 International Conference on Biomedical Engineering and Biotechnology The method achieved 82.45% balanced accuracy in distinguishing guilty from innocent subjects using reduced stimulus presentations, demonstrating that reliable P300-based lie detection could be implemented… [333] (2012)A novel protocol for detection of concealed information
20th Iranian Conference on Electrical Engineering (ICEE2012) Multimodal measurement combining ERP and autonomic signals achieved 90.9% classification accuracy using linear discriminant analysis, significantly exceeding single-modality approaches in laboratory CIT testing. [334] (2012)Concealed Knowledge Identification using facial thermal imaging
Li Creative Technologies The automated thermal imaging system achieved 83.5% accuracy in detecting concealed knowledge, comparable to or better than conventional polygraph testing (72–91% accuracy range). The non-contact… [335] (2011)Visual Cues for Facial Behaviour in Deception Detection
IEEE International Conference on Computer Applications and Industrial Electronics (ICCAIE) The pilot study demonstrated statistically supported potential for facial behaviour cues — including micro-expressions, facial asymmetry, lip biting, and blink rate changes — to discriminate… [336] (2011)Visual Cues of Facial Behaviour in Deception Detection
IEEE International Conference on Computer Applications and Industrial Electronics (ICCAIE) Human examiners achieved only 56.25% accuracy in detecting deception from facial cues — barely above chance — while a logistic regression classifier applied to the… [337] (2011)An investigation of data and text mining methods for real world deception detection
Expert Systems with Applications Information fusion ensemble models achieved >74% classification accuracy in detecting deception in authentic criminal investigation statements, outperforming artificial neural networks (73.46%) and decision trees (71.60%),… [338] (2011)Effects of Automated and Participative Decision Support in Computer-Aided Credibility Assessment
Journal of Management Information Systems Decision aids that collect both automated linguistic cues and human perceptual observations perform better than systems relying on only one type of cue, while participatory… [339] (2010)A New Computer Vision Based Polygraph System
Recent Patents on Computer Science The authors reported experimental validation of a computer vision system that successfully detected deception through automated pupil size measurement, claiming it represented the best computer… [340] (2010)Facial image processing for facial analysis
Conference proceedings Color-space pre-processing of visible images via Chrominance and Hue segmentation increased face detection rates and reduced false positives compared to standard Haar Classifiers, and enabled… [341] (2010)Commercialisation of an artificially intelligent deception detection system in the current security climate
International Conference on Fuzzy Systems The paper discusses commercialization pathways for AI-based deception detection in security contexts, highlighting that while the Silent Talker system achieved above-chance detection of deception versus… [342] (2010)Automatic blush detection in “Concealed information” test using visual stimuli
2010 International Conference of Soft Computing and Pattern Recognition The automated thermal detection system achieved approximately 77% True Positive Rate and 60% True Negative Rate in identifying participants with concealed information, demonstrating proof-of-concept but… [343] (2010)On Deception Detection in Multiagent Systems
IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans The correlation-based approach successfully detected deception in multi-agent systems by predicting agent responses and flagging statistically significant deviations, demonstrating that computational analysis of reasoning consistency… [344] (2010)Motion Profiles for Deception Detection Using Visual Cues
Lecture Notes in Computer Science The motion profile Nearest Neighbor models achieved 81.6% accuracy in detecting deception using only visual behavioral cues. Subject-specific models significantly outperformed general models, indicating that… [345] (2010)Detection of deception in structured interviews using sensors and algorithms
SPIE Proceedings Automated algorithms successfully detected deception from physiological features extracted during honest and deceptive interview intervals, with detection performance varying by interview style and reported through… [346] (2010)Technology Dominance in Complex Decision Making: The Case of Aided Credibility Assessment
Journal of Management Information Systems Both novices and professionals improved credibility assessment accuracy using the automated decision aid, but professionals showed no greater benefit than novices. Both groups frequently ignored… [347] (2009)Development of Technologies and Test Formats for Credibility Assessment
European Polygraph The paper likely presented validation data for novel technological approaches including thermal imaging systems and computerized scoring methods that could supplement or enhance traditional polygraph… [348] (2009)A new approach for EEG feature extraction in P300-based lie detection
Computer Methods and Programs in Biomedicine The study achieved 86% correct classification accuracy for both guilty and innocent subjects using a Genetic Algorithm-optimized multi-domain EEG feature set, surpassing the performance of… [349] (2009)Feature Selection for fMRI-based Deception Detection
BMC Bioinformatics Feature selection using an ensemble of filter and wrapper methods reduced the fMRI feature space from 65,166 to 124 voxels and significantly enhanced SVM classification… [350] (2009)Detecting concealment of intent in transportation screening: A proof of concept
IEEE Transactions on Intelligent Transportation Systems The research demonstrated that automated analysis of verbal and nonverbal behavior during interviews could detect concealed intent at above-chance levels, with kinesic tracking revealing differences… [351] (2009)Within-individual discrimination on the concealed information test using dynamic mixture modeling
Psychophysiology The hidden Markov discrimination method showed superior performance compared to conventional within-individual scoring approaches in a mock theft experiment with 34 participants, successfully modeling physiological… [352] (2009)The Preliminary Credibility Assessment System Embedded Algorithm Description and Validation Results
Johns Hopkins University Applied Physics Laboratory Validation studies showed 80% accuracy when inconclusives were excluded, with 23% inconclusive rates. When inconclusives were included, accuracy dropped to 63-79%. The algorithm was deliberately… [353] (2008)Periorbital Thermal Signal Extraction and Applications
30th Annual International IEEE EMBS Conference The fuzzy-based periorbital segmentation algorithm achieved an 82% deception classification success rate across four experimental runs, consistently outperforming the previous 10% hottest-pixels heuristic method by… [354] (2008)Brute-force comparison: A Monte Carlo study of the Objective Scoring System version 3 (OSS-3) and human polygraph scorers
Polygraph OSS-3 exceeded the decision accuracy of 10 human scorers and 9 out of 10 individual scorers across six accuracy dimensions. The algorithm demonstrated perfect reliability… [355] (2008)Multimodal Detection of Deception using Fusion of Reaction Time and P300 Component
2008 Cairo International Biomedical Engineering Conference The Bootstrapped Analysis of Reaction Time (BART) method achieved 81% accuracy, comparable to 80% for P300 wavelet analysis alone. The multimodal BART-BAD fusion method demonstrated… [356] (2008)The Language of Deception: Computational Approaches
Synthesis Lectures on Human Language Technologies Computational approaches reveal systematic linguistic differences between deception and truth, including distinctions between imaginative and informative language patterns, with practical applications across security, forensic, and… [357] (2008)Development and validation of the Objective Scoring System version 3
Polygraph OSS-3 achieved 88% average decision accuracy with balanced sensitivity and specificity, with computer algorithms providing perfect interrater reliability compared to human scorers (κ = .58–.61)… [358] (2008)User experiences with an unobtrusive decision aid for deception detection
2008 42nd Annual IEEE International Carnahan Conference on Security Technology Novice lie-catchers using the decision aid experienced improved deception detection accuracy, but users who incorporated more system recommendations performed better. Novices primarily turned to the… [359] (2008)Exploration of Feature Selection and Advanced Classification Models for High-Stakes Deception Detection
Proceedings of the 41st Annual Hawaii International Conference on System Sciences (HICSS 2008) The study achieved 76% accuracy in distinguishing truthful from deceptive statements, demonstrating that careful feature selection could improve classification performance while preventing model overfitting. [360] (2008)Video-Based Deception Detection
Studies in Computational Intelligence The automated system extracted behavioral features from video that performed comparably to human-coded perceptions in detecting deception, identifying behavioral profiles including agitated, over-controlled, and relaxed… [361] (2007)Imaging Facial Physiology for the Detection of Deceit
International Journal of Computer Vision The automated periorbital thermal imaging system achieved 87.2% overall deception classification accuracy across 39 subjects in a realistic mock-crime scenario, with detection rates of 87.5%… [362] (2007)Forehead Thermal Signature Extraction in Lie Detection
Proceedings of the 29th Annual International Conference of the IEEE EMBS The forehead thermal channel achieved a 76.3% success rate in classifying deceptive states in a 38-subject mock-crime sample, confirming the corrugator muscle's supraorbital vasculature as… [363] (2007)An Investigation of Heuristics of Human Judgment in Detecting Deception and Potential Implications in Countering Social Engineering
2007 IEEE Intelligence and Security Informatics An automated system using discriminant analysis significantly outperformed human judges in detecting deception, validating the potential of computational tools to overcome inherent human cognitive biases… [364] (2006)Improving a textual deception detection model
Proceedings of the 2006 conference of the Center for Advanced Studies on Collaborative research - CASCON '06 The study revealed that each of the four kinds of words in the Pennebaker model acts as a separate latent factor for deception, rather than… [365] (2006)Silent talker: a new computer-based system for the analysis of facial cues to deception
Applied Cognitive Psychology The Silent Talker system achieved 74% accuracy (p < 0.001) classifying individual deceptive responses and 80% accuracy (p = 0.035) for overall interview classifications in… [366] (2006)Temporal Context Lie Detection and Generation
Lecture Notes in Computer Science The research demonstrated that Dynamic Bayesian Networks can effectively model both the detection and generation of lies in temporal context systems by exploiting dependencies between… [367] (2006)Personality factors in human deception detection: comparing human to machine performance
Interspeech 2006 Human judges performed worse than chance and worse than automated machine learning systems when detecting deception from audio interviews. Personality factors measured by the NEO… [368] (2006 • P300 GKT COMPARISON)A Comparison of Methods for ERP Assessment in a P300-Based GKT
International Journal of Psychophysiology Compared three classifiers for P300-based detection achieving 74–80% correct detection rates. Demonstrated potential of ERP methods though accuracy remained below traditional polygraph. [369] (2005)Language of lies in prison: Linguistic classification of inmates' truthful and deceptive natural language
Applied Cognitive Psychology Automated linguistic analysis achieved 69.7-71.1% accuracy in classifying prisoner statements, with software showing unbiased criteria while human judges (especially younger prisoners) exhibited decision bias. Spatial… [370] (2005)Distinguishing Deceptive from Non-Deceptive Speech
Interspeech 2005 - Eurospeech Machine learning classifiers achieved approximately 70% accuracy in distinguishing truthful from deceptive speech, outperforming human judges by 12 percentage points. Contrary to conventional wisdom, filled… [371] (2005)Blob Analysis of the Head and Hands: A Method for Deception Detection
Proceedings of the 38th Annual Hawaii International Conference on System Sciences The study successfully validated automated blob analysis for tracking head and hand movements across diverse skin tones, demonstrating the technical feasibility of using computer vision… [372] (2005)An Empirical Investigation of Deception Behaviorin Instant Messaging
IEEE Transactions on Professional Communication Three types of nonverbal behaviors and three types of verbal behaviors significantly differentiated deceivers from truth tellers, including greater word count, more turn-taking, increased negative… [373] (2005)Using a linguistic analysis tool to detect deception
PsycEXTRA Dataset Technological tools showed promise for detecting deception through linguistic analysis, but effectiveness varied significantly based on communication medium and personal investment in the topic, demonstrating… [374] (2005 • MACHINE LEARNING)Classifying Spatial Patterns of Brain Activity with Machine Learning Methods: Application to Lie Detection
NeuroImage Applied machine learning to fMRI data achieving up to 100% within-subject and 88% cross-subject classification. Demonstrated potential of multivariate pattern analysis for deception detection. [375] (2004)Lie detection using thermal imaging
Thermosense XXVI (SPIE Proceedings) The study demonstrated that periorbital blood flow changes captured via mid-infrared thermal imaging can be computationally classified to distinguish deceptive from non-deceptive subjects, with the… [376] (2004)Detection of the Cognitive Components of Brain Potentials Using Wavelet Coefficients
Iranian Journal of Biomedical Engineering Discrete wavelet transform successfully decomposed single-trial P300 ERP signals into informative time-frequency features that distinguished guilty from innocent subjects, with subsequent work confirming 86% correct… [377] (2004)Computer-Aided Support of the Detection of Deception
Group Decision and Negotiation The research program demonstrated that automated detection systems achieved 70-92% accuracy compared to 54% for human judgment, with multimodal approaches combining kinesic, vocal, and linguistic… [378] (2004)Criminal Identity Deception and Deception Detection in Law Enforcement
Group Decision and Negotiation The study found that 30% of suspects used intentionally false identities, with 62.5% using partly deceptive names and 29.2% using completely false names. An algorithmic… [379] (2004)An exploratory study on promising cues in deception detection and application of decision tree
37th Annual Hawaii International Conference on System Sciences, 2004. Proceedings of the Five comparative experiments demonstrated that decision tree performance — including both accuracy rates and model complexity — was dramatically improved by statistically and semantically selecting… [380] (2004)The Influence of Task Load and Automation Trust on Deception Detection
Group Decision and Negotiation Participants under high task load exhibited over-reliance on automated systems regardless of their trust levels, while low task load conditions showed the expected trust-use relationship.… [381] (2004)Detection of deception: collaboration systems and technology
37th Annual Hawaii International Conference on System Sciences, 2004. Proceedings of the The paper established collaboration systems as platforms for embedded credibility assessment, framing research showing that technology-enabled multimodal analysis of verbal, vocal, and behavioral cues could… [382] (2003)Automated Scoring of Polygraph Data
Statistical Data Mining and Knowledge Discovery With a simplified logistic regression approach using carefully selected features from electrodermal, cardiovascular, and respiratory channels, the study achieved classification accuracy rates comparable to more… [383] (2003)Lying words: Predicting deception from linguistic styles
Personality and Social Psychology Bulletin The LIWC classification model achieved 67% accuracy when topic was constant and 61% overall accuracy across all studies—significantly better than chance (50%) and untrained human… [384] (2002)Thermal Facial Screening for Deception Detection
Proceedings of the Second Joint EMBS/BMES Conference Thermal facial imaging achieved an overall correct classification rate of 83% in a mock-crime paradigm (versus 70% for traditional polygraph by expert examiners), with 90%… [385] (2002)Human behaviour: Seeing through the face of deception
Nature Thermal imaging of the periorbital region achieved 84% classification accuracy in detecting deception during mock crime interrogations, with performance comparable to traditional polygraph examination, though… [386] (2002)Computerized Polygraph System
Journal of Forensic Sciences The CPS algorithm demonstrated accuracy rates of 95-96% on confirmed truthful cases and 85-95% on deceptive cases in field validation studies, performing comparably to or… [387] (2001)Nonlinear Feature Based Classification of Speech under Stress
IEEE Transactions on Speech and Audio Processing The TEO-CB-Auto-Env feature substantially outperformed traditional pitch and MFCC features, achieving consistent classification rates of 73.9%–85.2% across different stress styles and maintaining performance in both… [388] (2000)Thermal imaging for anxiety detection
Proceedings 2000 IEEE Workshop on Computer Vision Beyond the Visible Spectrum The study demonstrated proof-of-concept feasibility for using thermal image analysis to detect facial anxiety patterns at a distance without physical sensor contact, with measurable temperature… [389] (1999)Measuring facial expressions by computer image analysis
Psychophysiology The automated hybrid system achieved 91% accuracy in recognizing six upper facial Action Units, outperformed non-expert humans, matched expert-level performance, and reduced coding time from… [390] (1993)The Use of Fuzzy Set Classification for Pattern Recognition of the Polygraph
The fuzzy k-nearest neighbor algorithm successfully classified polygraph data using optimized feature combinations from multiple physiological channels, demonstrating that machine learning approaches could extract meaningful… [391] (1993)Multivariate classifiers perform as well as experts in the detection of deception
Statistical multivariate classifiers — including discriminant analysis and a bootstrap approach — achieved accuracy statistically equivalent to expert human evaluators on polygraph CQT data, with… [392] (1992)Bootstrap Decision Making for Polygraph Examinations,
Bootstrap resampling outperformed both human numerical evaluation and discriminant analysis in absolute terms. However, the expert human scorers likely overestimated typical field examiner accuracy, making… [393] (1992)Neural network classifiers and the detection of deception revisited: Depth of learning and overfitting
Psychophysiology Neural network classifiers applied to polygraph data were found to be highly susceptible to overfitting, with deeper training producing inflated apparent accuracy that sharply deteriorated… [394] (1990)Development of a Computerized Polygraph System and Physiological Measures for Detection of Deception and Countermeasures: A Pilot Study
Contract Report The study demonstrated the feasibility of fully computerized polygraph systems for both data acquisition and automated analysis, contributing to algorithms and measurement techniques that formed… [395] (1982)Cross-validation of a computerized diagnostic procedure for detection of deception
Psychophysiology Computerized discriminant analysis achieved accuracy rates equivalent to experienced human evaluators in classifying guilty and innocent subjects, with successful cross-validation demonstrating generalizability across independent samples. [396] (0)Coalitional Tracker for Deception Detection in Thermal Imagery
Augmented Vision Perception in Infrared The coalitional tracking method achieved 82% deception classification accuracy, a 20% improvement over previous methods, while maintaining pixel-level tracking accuracy even during unrestricted subject movement.… [397] (0)Modality Effects in Deception Detection and Applications in Automatic-Deception-Detection
Proceedings of the 38th Annual Hawaii International Conference on System Sciences Differences between deception and truth were generally consistent across the three modalities, with only specificity showing significant modality-by-condition interaction effects. Modality had strong main effects…Related research topics
Explore more research.
Search and filter our complete database of 3493 peer-reviewed studies.