NeuroHybridNet: Otokodlayıcı Destekli Fraktal-Dalgaçık Özelliklerle Zenginleştirilmiş Hibrit Bir EEG Sınıflandırma Modeli

Öz

Elektroensefalografi (EEG) sinyalleri, özellikle sürücü dikkatinin izlenmesi gibi dinamik ve yüksek riskli ortamlarda bilişsel durumların değerlendirilmesi açısından zengin ancak karmaşık bir bilgi kaynağı sunar. Bu çalışmada, WaveFrac-AE adı verilen, zaman-frekans analizi, fraktal dinamikler ve doğrusal olmayan betimleyicileri Otokodlayıcı (Autoencoder) tabanlı boyut indirgeme ile birleştiren hibrit bir EEG sınıflandırma modeli önerilmektedir. Dikkatli, dikkati dağılmış ve uykulu olmak üzere üç bilişsel durumu temsil eden EEG sinyalleri, hem zamansal hem de yapısal karmaşıklığı yansıtan kompakt bir gizil uzaya dönüştürülmektedir. Model, Sürekli Dalgaçık Dönüşümü (CWT) özelliklerini, fraktal ölçütlerle (Petrosian Fraktal Boyutu, Hurst Üssü, DFA) birleştirerek EEG sinyallerinin durağan olmayan doğasını sağlam bir biçimde temsil eder. Otokodlayıcı bileşeni, gürültü ve gereksiz bilgileri filtreleyerek genellenebilirliği artırır ve gerçek zamanlı nöroergonomik uygulamalar için doğru ve ölçeklenebilir sınıflandırma sağlar. Bu hibrit yaklaşım, XGBoost, LightGBM, CatBoost, Random Forest ve SVM gibi çeşitli makine öğrenmesi algoritmalarıyla test edilmiştir. Kişiye özel analizlerde SVM ortalama %97’nin üzerinde doğruluk sağlarken, tüm deneklerin birleşik veri setinde %92’nin üzerinde doğruluk elde edilmiştir. Güncel çalışmalarla yapılan karşılaştırmalar, yöntemin rekabetçi olduğunu ve birçok durumda üstün performans sergilediğini göstermektedir. Sonuç olarak, önerilen model, sürücü dikkat takibi sistemleri için yüksek performanslı bir çözüm sunmakta ve akıllı araçlara entegre erken uyarı sistemlerinin geliştirilmesi açısından önemli bir potansiyele sahiptir.

Anahtar Kelimeler

• Otokodlayıcı
• Sürekli Dalgaçık Dönüşümü (CWT)
• EEG
• Fraktal Boyut
• Makine Öğrenmesi

NeuroHybridNet: A Hybrid EEG Classification Model with Autoencoder-Enhanced Fractal-Wavelet Features

Abstract

Electroencephalography (EEG) signals offer a rich but complex source of information for assessing cognitive states, particularly in dynamic and high-stakes environments such as driver attention monitoring, where rapid and accurate detection of mental fatigue or distraction is critical for safety and performance.This study proposes WaveFrac-AE, a hybrid EEG classification model that combines time-frequency analysis, fractal dynamics, and nonlinear descriptors with Autoencoder-based dimensionality reduction. EEG signals representing three cognitive states—focused, distracted, and drowsy—are transformed into a compact latent space capturing both temporal and structural complexity. By fusing Continuous Wavelet Transform features with fractal measures (Petrosian, Hurst, DFA), the model achieves robust representation of non-stationary EEG dynamics. The Autoencoder component enhances generalizability by filtering noise and redundancy, enabling accurate and scalable mental state classification for real-time neuroergonomic applications. This hybrid approach has been tested with various machine learning algorithms—namely XGBoost, LightGBM, CatBoost, Random Forest, and Support Vector Machines (SVM). In subject-specific analyses, SVM achieved an average accuracy exceeding 97%, while the aggregated dataset—combining all subjects—yielded an accuracy surpassing 92%. Comparisons with contemporary studies suggest that this method occupies a competitive position and, in numerous cases, demonstrates higher performance. Consequently, the proposed model offers high-performance solution in driver attentiveness monitoring systems, thereby showing substantial potential for the development of early warning systems integrated into smart automobiles.

Keywords

• Autoencoder
• CWT
• EEG
• Fractal Dimension
• Machine Learning.

References

1. [1] M. Yousaf, M. Farhan, Y. Saeed, M. J. Iqbal, F. Ullah, and G. Srivastava, “Enhancing driver attention and road safety through EEG-informed deep reinforcement learning and soft computing,” Appl Soft Comput, vol. 167, p. 112320, Dec. 2024, doi: 10.1016/J.ASOC.2024.112320.
2. [2] T. Zhang, J. Yang, M. Chen, Z. Li, J. Zang, and X. Qu, “EEG-based assessment of driver trust in automated vehicles,” Expert Syst Appl, vol. 246, p. 123196, Jul. 2024, doi: 10.1016/J.ESWA.2024.123196.
3. [3] J. Min, C. Xiong, Y. Zhang, and M. Cai, “Driver fatigue detection based on prefrontal EEG using multi-entropy measures and hybrid model,” Biomed Signal Process Control, vol. 69, p. 102857, Aug. 2021, doi: 10.1016/J.BSPC.2021.102857.
4. [4] A. Khosla, P. Khandnor, and T. Chand, “Automated diagnosis of depression from EEG signals using traditional and deep learning approaches: A comparative analysis,” Biocybern Biomed Eng, vol. 42, no. 1, pp. 108–142, Jan. 2022, doi: 10.1016/J.BBE.2021.12.005.
5. [5] Y. Xie and S. Oniga, “Classification of Motor Imagery EEG Signals Based on Data Augmentation and Convolutional Neural Networks,” Sensors 2023, Vol. 23, Page 1932, vol. 23, no. 4, p. 1932, Feb. 2023, doi: 10.3390/S23041932.
6. [6] S. Morales and M. E. Bowers, “Time-frequency analysis methods and their application in developmental EEG data,” Dev Cogn Neurosci, vol. 54, p. 101067, Apr. 2022, doi: 10.1016/J.DCN.2022.101067.
7. [7] C. Uyulan and T. T. Erguzel, “Analysis of Time — Frequency EEG Feature Extraction Methods for Mental Task Classification,” International Journal of Computational Intelligence Systems, vol. 10, no. 1, pp. 1280–1288, Jan. 2017, doi: 10.2991/IJCIS.10.1.87/METRICS.
8. [8] H. Steiner, I. Mikheev, and O. Martynova, “Cross-Subject Classification of Effectiveness in Performing Cognitive Tasks Using Resting-State EEG,” Applied Sciences 2023, Vol. 13, Page 6606, vol. 13, no. 11, p. 6606, May 2023, doi: 10.3390/APP13116606.

9. [9] E. Guttmann-Flury, X. Sheng, and X. Zhu, “Channel selection from source localization: A review of four EEG-based brain–computer interfaces paradigms,” Behav Res Methods, vol. 55, no. 4, pp. 1980–2003, Jun. 2023, doi: 10.3758/S13428-022-01897-2/METRICS.
10. [10] E. Gibson, N. J. Lobaugh, S. Joordens, and A. R. McIntosh, “EEG variability: Task-driven or subject-driven signal of interest?,” Neuroimage, vol. 252, p. 119034, May 2022, doi: 10.1016/J.NEUROIMAGE.2022.119034.
11. [11] Y. Hadad, M. Bensimon, Y. Ben-Shimol, and S. Greenberg, “Situational Awareness Classification Based on EEG Signals and Spiking Neural Network,” Applied Sciences 2024, Vol. 14, Page 8911, vol. 14, no. 19, p. 8911, Oct. 2024, doi: 10.3390/APP14198911.
12. [12] A. Al-Nafjan and M. Aldayel, “Predict Students’ Attention in Online Learning Using EEG Data,” Sustainability 2022, Vol. 14, Page 6553, vol. 14, no. 11, p. 6553, May 2022, doi: 10.3390/SU14116553.
13. [13] S. Morales and M. E. Bowers, “Time-frequency analysis methods and their application in developmental EEG data,” Dev Cogn Neurosci, vol. 54, p. 101067, Apr. 2022, doi: 10.1016/J.DCN.2022.101067.
14. [14] Ç. İ. Acı, M. Kaya, and Y. Mishchenko, “Distinguishing mental attention states of humans via an EEG-based passive BCI using machine learning methods,” Expert Syst Appl, vol. 134, pp. 153–166, Nov. 2019, doi: 10.1016/J.ESWA.2019.05.057.
15. [15] I. Alreshidi, I. Moulitsas, and K. W. Jenkins, “Multimodal Approach for Pilot Mental State Detection Based on EEG,” Sensors 2023, Vol. 23, Page 7350, vol. 23, no. 17, p. 7350, Aug. 2023, doi: 10.3390/S23177350.
16. [16] J. K. Nuamah and Y. Seong, “Support vector machine (SVM) classification of cognitive tasks based on electroencephalography (EEG) engagement index,” Brain-Computer Interfaces, vol. 5, no. 1, pp. 1–12, Jan. 2018, doi: 10.1080/2326263X.2017.1338012.
17. [17] A. Hag, D. Handayani, T. Pillai, T. Mantoro, M. H. Kit, and F. Al-Shargie, “EEG Mental Stress Assessment Using Hybrid Multi-Domain Feature Sets of Functional Connectivity Network and Time-Frequency Features,” Sensors 2021, Vol. 21, Page 6300, vol. 21, no. 18, p. 6300, Sep. 2021, doi: 10.3390/S21186300.
18. [18] A. A. Rahman et al., “Detection of Mental State from EEG Signal Data: An Investigation with Machine Learning Classifiers,” KST 2022 - 2022 14th International Conference on Knowledge and Smart Technology, pp. 152–166, 2022, doi: 10.1109/KST53302.2022.9729084.
19. [19] H. Zeng et al., “An EEG-Based Transfer Learning Method for Cross-Subject Fatigue Mental State Prediction,” Sensors 2021, Vol. 21, Page 2369, vol. 21, no. 7, p. 2369, Mar. 2021, doi: 10.3390/S21072369.
20. [20] N. V. Kimmatkar and B. Vijaya Babu, “Novel Approach for Emotion Detection and Stabilizing Mental State by Using Machine Learning Techniques,” Computers 2021, Vol. 10, Page 37, vol. 10, no. 3, p. 37, Mar. 2021, doi: 10.3390/COMPUTERS10030037.
21. [21] L. D. Sharma, V. K. Bohat, M. Habib, A. M. Al-Zoubi, H. Faris, and I. Aljarah, “Evolutionary inspired approach for mental stress detection using EEG signal,” Expert Syst Appl, vol. 197, p. 116634, Jul. 2022, doi: 10.1016/J.ESWA.2022.116634.
22. [22] D. Kamińska, K. Smółka, and G. Zwoliński, “Detection of Mental Stress through EEG Signal in Virtual Reality Environment,” Electronics 2021, Vol. 10, Page 2840, vol. 10, no. 22, p. 2840, Nov. 2021, doi: 10.3390/ELECTRONICS10222840.
23. [23] V. G. Rajendran, S. Jayalalitha, K. Adalarasu, and R. Mathi, “Machine learning based human mental state classification using wavelet packet decomposition-an EEG study,” Multimed Tools Appl, vol. 83, no. 35, pp. 83093–83112, Oct. 2024, doi: 10.1007/S11042-024-18725-8/TABLES/8.
24. [24] O. Faust, R. Acharya U, S. M. Krishnan, and L. C. Min, “Analysis of cardiac signals using spatial filling index and time-frequency domain,” Biomed Eng Online, vol. 3, no. 1, pp. 1–11, Sep. 2004, doi: 10.1186/1475-925X-3-30/FIGURES/8.
25. [25] P. Ghorbanian, D. M. Devilbiss, T. Hess, A. Bernstein, A. J. Simon, and H. Ashrafiuon, “Exploration of EEG features of Alzheimer’s disease using continuous wavelet transform,” Med Biol Eng Comput, vol. 53, no. 9, pp. 843–855, Sep. 2015, doi: 10.1007/S11517-015-1298-3/FIGURES/4.
26. [26] A. K. Maddirala and K. C. Veluvolu, “SSA with CWT and k-Means for Eye-Blink Artifact Removal from Single-Channel EEG Signals,” Sensors 2022, Vol. 22, Page 931, vol. 22, no. 3, p. 931, Jan. 2022, doi: 10.3390/S22030931.
27. [27] A. M. John, O. Elfanagely, C. A. Ayala, M. Cohen, and C. J. Prestigiacomo, “The utility of fractal analysis in clinical neuroscience,” Rev Neurosci, vol. 26, no. 6, pp. 633–645, Dec. 2015, doi: 10.1515/REVNEURO-2015-0011/MACHINEREADABLECITATION/RIS.
28. [28] S. Morales and M. E. Bowers, “Time-frequency analysis methods and their application in developmental EEG data,” Dev Cogn Neurosci, vol. 54, p. 101067, Apr. 2022, doi: 10.1016/J.DCN.2022.101067.
29. [29] M. Pal, P. Manimaran, and P. K. Panigrahi, “A multi scale time–frequency analysis on Electroencephalogram signals,” Physica A: Statistical Mechanics and its Applications, vol. 586, p. 126516, Jan. 2022, doi: 10.1016/J.PHYSA.2021.126516.
30. [30] M. A. Rahman, F. Khanam, M. Ahmad, and M. S. Uddin, “Multiclass EEG signal classification utilizing Rényi min-entropy-based feature selection from wavelet packet transformation,” Brain Inform, vol. 7, no. 1, pp. 1–11, Dec. 2020, doi: 10.1186/S40708-020-00108-Y/TABLES/3.
31. [31] U. R. Acharya, H. Fujita, V. K. Sudarshan, S. Bhat, and J. E. W. Koh, “Application of entropies for automated diagnosis of epilepsy using EEG signals: A review,” Knowl Based Syst, vol. 88, pp. 85–96, Nov. 2015, doi: 10.1016/J.KNOSYS.2015.08.004.
32. [32] P. Patel, R. Raghunandan, and R. N. Annavarapu, “EEG-based human emotion recognition using entropy as a feature extraction measure,” Brain Inform, vol. 8, no. 1, pp. 1–13, Dec. 2021, doi: 10.1186/S40708-021-00141-5/FIGURES/4.
33. [33] M. A. Belcher, I. C. Hwang, S. Bhattacharya, W. D. Hairston, and J. S. Metcalfe, “EEG-based prediction of driving events from passenger cognitive state using Morlet Wavelet and Evoked Responses,” Transportation Engineering, vol. 8, p. 100107, Jun. 2022, doi: 10.1016/J.TRENG.2022.100107.
34. [34] S. A. David, J. A. T. Machado, C. M. C. Inácio, and C. A. Valentim, “A combined measure to differentiate EEG signals using fractal dimension and MFDFA-Hurst,” Commun Nonlinear Sci Numer Simul, vol. 84, p. 105170, May 2020, doi: 10.1016/J.CNSNS.2020.105170.
35. [35] P. Chattopadhyay and P. Konar, “Feature Extraction using Wavelet Transform for Multi-class Fault Detection of Induction Motor,” Journal of The Institution of Engineers (India): Series B, vol. 95, no. 1, pp. 73–81, Jan. 2014, doi: 10.1007/S40031-014-0076-1/TABLES/11.
36. [36] A. F. Farag, S. M. El-Metwally, and A. A. A. Morsy, “Automated Sleep Staging Using Detrended Fluctuation Analysis of Sleep EEG,” Advances in Intelligent Systems and Computing, vol. 195 AISC, pp. 501–510, 2013, doi: 10.1007/978-3-642-33941-7_44.
37. [37] Y. Wang, H. Yao, and S. Zhao, “Auto-encoder based dimensionality reduction,” Neurocomputing, vol. 184, pp. 232–242, Apr. 2016, doi: 10.1016/J.NEUCOM.2015.08.104.
38. [38] Y. Liu, B. Jiang, J. Feng, J. Hu, and H. Zhang, “Classification of EEG Signals for Epileptic Seizures Using Feature Dimension Reduction Algorithm based on LPP,” Multimed Tools Appl, vol. 80, no. 20, pp. 30261–30282, Aug. 2021, doi: 10.1007/S11042-020-09135-7/FIGURES/10.
39. [39] R. Sharma and H. K. Meena, “Emerging Trends in EEG Signal Processing: A Systematic Review,” SN Comput Sci, vol. 5, no. 4, pp. 1–14, Apr. 2024, doi: 10.1007/S42979-024-02773-W/TABLES/5.
40. [40] R. Mandal, S. Pal, and U. Maji, “Importance of Activity and Emotion Detection in the Field of Ambient Assisted Living,” Studies in Computational Intelligence, vol. 1175, pp. 209–240, 2024, doi: 10.1007/978-3-031-71821-2_7.
41. [41] S. K. Kiangala and Z. Wang, “An effective adaptive customization framework for small manufacturing plants using extreme gradient boosting-XGBoost and random forest ensemble learning algorithms in an Industry 4.0 environment,” Machine Learning with Applications, vol. 4, p. 100024, Jun. 2021, doi: 10.1016/J.MLWA.2021.100024.
42. [42] B. L. Radhakrishnan, K. Ezra, I. J. Jebadurai, I. Selvakumar, and P. Karthikeyan, “An Autonomous Sleep-Stage Detection Technique in Disruptive Technology Environment,” Sensors, vol. 24, no. 4, p. 1197, Feb. 2024, doi: 10.3390/S24041197/S1.
43. [43] S. Singh, H. Jadli, R. Padma Priya, and V. B. Surya Prasath, “KDTL: knowledge-distilled transfer learning framework for diagnosing mental disorders using EEG spectrograms,” Neural Comput Appl, vol. 36, no. 30, pp. 18919–18934, Oct. 2024, doi: 10.1007/S00521-024-10207-0/TABLES/5.
44. [44] T. O. Omotehinwa, D. O. Oyewola, and E. G. Moung, “Optimizing the light gradient-boosting machine algorithm for an efficient early detection of coronary heart disease,” Informatics and Health, vol. 1, no. 2, pp. 70–81, Sep. 2024, doi: 10.1016/J.INFOH.2024.06.001.
45. [45] B. K. Swain, S. Mohapatra, M. Mishra, and R. Sharma, “A unified approach for Parkinson’s disease recognition: imbalance mitigation and grid search optimized boosting with LightGBM,” Med Biol Eng Comput, vol. 62, no. 11, pp. 3471–3491, Nov. 2024, doi: 10.1007/S11517-024-03139-3/TABLES/12.
46. [46] T. O. Omotehinwa, D. O. Oyewola, and E. G. Moung, “Optimizing the light gradient-boosting machine algorithm for an efficient early detection of coronary heart disease,” Informatics and Health, vol. 1, no. 2, pp. 70–81, Sep. 2024, doi: 10.1016/J.INFOH.2024.06.001.
47. [47] S. K. Pemmada, J. Nayak, and A. R. Routray, “A Photoplethysmography Based Mental Workload Evaluation Using Ensembled CatBoost Approach,” Journal of The Institution of Engineers (India): Series B, vol. 106, no. 1, pp. 165–180, Jun. 2024, doi: 10.1007/S40031-024-01092-1/FIGURES/12.
48. [48] M. Bałdyga, K. Barański, J. Belter, M. Kalinowski, and P. Weichbroth, “Anomaly Detection in Railway Sensor Data Environments: State-of-the-Art Methods and Empirical Performance Evaluation,” Sensors 2024, Vol. 24, Page 2633, vol. 24, no. 8, p. 2633, Apr. 2024, doi: 10.3390/S24082633.
49. [49] P. Zhang, X. Wang, W. Zhang, and J. Chen, “Learning Spatial-Spectral-Temporal EEG Features With Recurrent 3D Convolutional Neural Networks for Cross-Task Mental Workload Assessment,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 27, no. 1, pp. 31–42, Jan. 2019, doi: 10.1109/TNSRE.2018.2884641.
50. [50] E. Q. Wu, X. Y. Peng, C. Z. Zhang, J. X. Lin, and R. S. F. Sheng, “Pilots’ fatigue status recognition using deep contractive autoencoder network,” IEEE Trans Instrum Meas, vol. 68, no. 10, pp. 3907–3919, Oct. 2019, doi: 10.1109/TIM.2018.2885608.
51. [51] X. Li, J. Tang, X. Li, and Y. Yang, “CWSTR-Net: A Channel-Weighted Spatial–Temporal Residual Network based on nonsmooth nonnegative matrix factorization for fatigue detection using EEG signals,” Biomed Signal Process Control, vol. 97, p. 106685, Nov. 2024, doi: 10.1016/J.BSPC.2024.106685.
52. [52] F. Wang, S. Wu, W. Zhang, Z. Xu, Y. Zhang, and H. Chu, “Multiple nonlinear features fusion based driving fatigue detection,” Biomed Signal Process Control, vol. 62, p. 102075, Sep. 2020, doi: 10.1016/J.BSPC.2020.102075.
53. [53] H. Zeng et al., “An EEG-Based Transfer Learning Method for Cross-Subject Fatigue Mental State Prediction,” Sensors 2021, Vol. 21, Page 2369, vol. 21, no. 7, p. 2369, Mar. 2021, doi: 10.3390/S21072369.
54. [54] A. Sonnleitner et al., “EEG alpha spindles and prolonged brake reaction times during auditory distraction in an on-road driving study,” Accid Anal Prev, vol. 62, pp. 110–118, Jan. 2014, doi: 10.1016/J.AAP.2013.08.026.
55. [55] J. Cao, E. M. Garro, and Y. Zhao, “EEG/fNIRS Based Workload Classification Using Functional Brain Connectivity and Machine Learning,” Sensors, vol. 22, no. 19, p. 7623, Oct. 2022, doi: 10.3390/S22197623/S1.
56. [56] V. P. B and S. Chinara, “Automatic classification methods for detecting drowsiness using wavelet packet transform extracted time-domain features from single-channel EEG signal,” J Neurosci Methods, vol. 347, p. 108927, Jan. 2021, doi: 10.1016/J.JNEUMETH.2020.108927