Wavelet Decomposition and Learning Architectures for EEG-Based Lie Detection: A LOSOCV-Based Comparative Evaluation of DWT Feature Representations on the LieWaves Dataset

Abstract

This study systematically evaluates discrete wavelet transform (DWT)-based EEG feature representations and machine/deep learning classifiers for subject-independent deception detection using the publicly available LieWaves dataset. EEG signals were bandpass filtered, processed with Automatic and Tunable Artifact Removal (ATAR), segmented with the Overlapping Sliding Window (OSW) method, and decomposed using DWT. Four configurations (db4-Level4, db4-Level5, db5-Level4, and db5-Level5) were compared within an outer Leave-One-Subject-Out Cross-Validation (LOSOCV) framework using a three-subject, subject-wise inner holdout procedure; db5-Level5 achieved the highest mean inner-validation performance. Using the fixed db5-Level5 representation, DWT-derived statistical features were evaluated with an EEGNet-inspired compact convolutional model, CNN, LSTM, BiLSTM, GRU, Random Forest, KNN, and SVM under LOSOCV. The EEGNet-inspired model achieved the highest average performance, with an accuracy of 0.5779±0.0903 and a macro F1-score of 0.5295±0.1359. Nevertheless, all models showed limited subject-independent performance. These results highlight the importance of subject-independent validation for highly overlapping EEG segments and indicate that the proposed pipeline should be regarded as an exploratory comparative framework rather than a ready-to-use practical or forensic lie detection system.

Keywords

• EEG
• Lie Detection
• Machine Learning
• Deep Learning
• Discrete Wavelet Transform

Supporting Institution

This research received no external funding.

Ethical Statement

This study was conducted using publicly available datasets. No human participants or animals were directly involved in this research. Therefore, ethical approval was not required. The authors declare that they have no conflict of interest.

Thanks

This study was conducted within the scope of the Design Project course at the Department of Computer Engineering, Bilecik Şeyh Edebali University. The authors would like to express their gratitude to Aslan et al. for making the LieWaves dataset publicly available, which significantly contributed to the experimental evaluation of this research.

EEG Tabanlı Yalan Algılamada Dalgacık Ayrıştırma ve Öğrenme Mimarileri: LieWaves Veri Kümesi Üzerinde DWT Özellik Temsillerinin LOSOCV Tabanlı Karşılaştırmalı Değerlendirmesi

Abstract

Bu çalışma, kamuya açık LieWaves veri setini kullanarak denek bağımsız EEG tabanlı yalan tespiti için Ayrık Dalgacık Dönüşümü (DWT) tabanlı özellik temsillerini ve makine/derin öğrenme sınıflandırıcılarını sistematik olarak değerlendirmektedir. EEG sinyalleri bant geçiren filtreleme ve Otomatik ve Ayarlanabilir Artefakt Giderme (ATAR) işlemlerinden geçirilmiş, Örtüşen Kayar Pencere (OSW) yöntemiyle bölütlenmiş ve DWT ile ayrıştırılmıştır. Dört DWT konfigürasyonu (db4-Level4, db4-Level5, db5-Level4 ve db5-Level5), dış Bir Denek Dışarıda Bırakma Çapraz Doğrulama (LOSOCV) çerçevesinde üç denekli, denek bazlı iç holdout doğrulaması kullanılarak karşılaştırılmış; db5-Level5 en yüksek ortalama iç doğrulama performansına ulaşmıştır. Sabit db5-Level5 temsili kullanılarak DWT tabanlı istatistiksel özellikler EEGNet'ten esinlenen kompakt konvolüsyonel model, CNN, LSTM, BiLSTM, GRU, Rastgele Orman, KNN ve SVM ile LOSOCV altında değerlendirilmiştir. EEGNet'ten esinlenen model 0,5779±0,0903 doğruluk ve 0,5295±0,1359 makro F1 puanı ile en yüksek ortalama performansı elde etmiştir. Bununla birlikte, tüm modellerin denek bağımsız performansı sınırlı kalmıştır. Sonuçlar, yüksek ölçüde örtüşen EEG bölütlerinde denek bağımsız doğrulamanın önemini göstermekte ve önerilen iş akışının kullanıma hazır pratik/adli bir sistemden ziyade keşif amaçlı karşılaştırmalı bir çerçeve olarak değerlendirilmesi gerektiğini ortaya koymaktadır.

Keywords

• EEG
• Yalan Tespiti
• Makine Öğrenmesi
• Derin Öğrenme
• Ayrık Dalgacık Dönüşümü

Supporting Institution

Bu araştırma herhangi bir dış finansman almamıştır.

Ethical Statement

Bu çalışma, kamuya açık veri setleri kullanılarak gerçekleştirilmiştir. Bu araştırmaya doğrudan insan veya hayvan katılımcı dahil edilmemiştir. Bu nedenle, etik onay gerekmemiştir. Yazarlar, herhangi bir çıkar çatışması olmadığını beyan ederler.

Thanks

Bu çalışma, Bilecik Şeyh Edebali Üniversitesi Bilgisayar Mühendisliği Bölümü Tasarım Projesi dersi kapsamında gerçekleştirilmiştir. Yazarlar, bu araştırmanın deneysel değerlendirmesine önemli katkı sağlayan LieWaves veri setini kamuya açık hale getiren Aslan ve arkadaşlarına teşekkürlerini sunarlar.

References

1. Acharya, U.R., Oh, S.L., Hagiwara, Y., Tan, J.H., Adeli, H., 2018. Deep convolutional neural network for the automated detection of seizures using EEG signals. Computers in Biology and Medicine, 100, 270-278.
2. Amin, H.U., Mumtaz, W., Subhani, A.R., Saad, M.N.M., Malik, A.S., 2017. Classification of EEG signals based on pattern recognition approach. Frontiers in Computational Neuroscience, 11, 103.
3. Aslan, M., Baykara, M., Alakus, T.B., 2024a. LieWaves: dataset for lie detection based on EEG signals and wavelets. Medical & Biological Engineering & Computing, 62(5), 1571-1588.
4. Aslan, M., Baykara, M., Alakus, T.B., 2024b. LSTMNCP: lie detection from EEG signals with novel hybrid deep learning method. Multimedia Tools and Applications, 83(11), 31655-31671.
5. Bhatt, P., Sethi, A., Tasgaonkar, V., Shroff, J., Pendharkar, I., Desai, A., Jain, N.K., 2023. Machine learning for cognitive behavioral analysis: datasets, methods, paradigms, and research directions. Brain Informatics, 10(1), 18.
6. Chanda, D., Mandal, R.K., 2023. A comprehensive review and future prospects of lie detection using machine learning. International Conference on Computational Technologies and Electronics, November 2023, pp. 64-77.
7. Ekman, P., 2009. Telling lies: Clues to deceit in the marketplace, politics, and marriage, 4th ed., W.W. Norton & Company, New York.
8. Farwell, L.A., Donchin, E., 1991. The truth will out: Interrogative polygraphy (“lie detection”) with event-related brain potentials. Psychophysiology, 28(5), 531-547.

9. Joshi, G., Tasgaonkar, V., et al., 2025. Multimodal machine learning for deception detection using behavioral and physiological data. Scientific Reports, 15(1), 8943.
10. Lawhern, V.J., Solon, A.J., Waytowich, N.R., Gordon, S.M., Hung, C.P., Lance, B.J., 2018. EEGNet: A compact convolutional neural network for EEG-based brain-computer interfaces. Journal of Neural Engineering, 15(5), 056013.
11. Nagale, T., Khandare, A., 2025. Spontaneous deception detection using EEG signals: Advances, challenges, and future directions in deep learning frameworks. 2025 IEEE International Conference on Emerging Trends in Engineering and Computing (ETECOM), October 2025, pp. 1-6.
12. Pérez-Rosas, V., Abouelenien, M., Mihalcea, R., Burzo, M., 2015. Deception detection using real-life trial data. Proceedings of the 2015 ACM International Conference on Multimodal Interaction, Seattle, USA, November 2015, pp. 59-66.
13. Rahmani, M., Mohajelin, F., Khaleghi, N., Sheykhivand, S., Danishvar, S., 2024. An automatic lie detection model using EEG signals based on the combination of type 2 fuzzy sets and deep graph convolutional networks. Sensors, 24(11), 3598.
14. Raj, V.A., Parupudi, T., Thalengala, A., Nayak, S.G., 2025. A comprehensive review of deep learning models for denoising EEG signals: challenges, advances, and future directions. Discover Applied Sciences, 7(11), 1268.
15. Rosenfeld, J.P., Soskins, M., Bosh, G., Ryan, A., 2004. Simple, effective countermeasures to P300-based tests of detection of concealed information. Psychophysiology, 41(2), 205-219.
16. Roy, Y., Banville, H., Albuquerque, I., Gramfort, A., Falk, T.H., Faubert, J., 2019. Deep learning-based electroencephalography analysis: a systematic review. Journal of Neural Engineering, 16(5), 051001.
17. Şen, M.U., Perez-Rosas, V., Yanikoglu, B., Abouelenien, M., Burzo, M., Mihalcea, R., 2022. Multimodal deception detection using real-life trial data. IEEE Transactions on Affective Computing, 13(1), 306-319.
18. Subasi, A. (2007). EEG signal classification using wavelet feature extraction and a mixture of expert model. Expert Systems with Applications, 32(4), 1084-1093.
19. Taha, B.N., Baykara, M., Alakuş, T.B., 2025. Neurophysiological approaches to lie detection: A systematic review. Brain Sciences, 15(5), 519.
20. Tarawneh, A.A., Kamel, A., Naceur, M.S., 2025. Comparative performance analysis of machine and deep learning models for EEG-based biometric authentication. Informatica, 49(8).
21. Uotinen, M., Simola, P., Henttonen, P.J., 2024. Detecting deceit within a predominantly true statement using two parallel assessment methods: A pilot study. Journal of Military Studies, 1(1).
22. Vrij, A., 2008. Detecting lies and deceit: Pitfalls and opportunities, 2nd ed., John Wiley & Sons, Chichester.
23. Vrij, A., Granhag, P.A., Porter, S., 2010. Pitfalls and opportunities in nonverbal and verbal lie detection. Psychological Science in the Public Interest, 11(3), 89-121.