Journal Press India^®

Author Details ( * ) denotes Corresponding author

We present a Hybrid Diffusion–Transformer Model for EEG-based prognosis of epilepsy, designed to capture both the noisy generative structure of EEG signals and the long-range temporal dependencies essential for clinical prediction. The model begins by dividing multichannel EEG recordings into short epochs and extracting spectral–temporal feature vectors that summarize activity across major brain rhythms, including delta, theta, alpha, beta, and gamma bands. These features form a structured sequence that reflects neural fluctuations relevant to disease progression. To address the inherent noise and variability of EEG data, a diffusion-based generative prior learns how latent EEG patterns degrade under noise and how clean representations can be recovered through a learned denoising process. The resulting latent embedding is combined with the observed rhythm features to strengthen the model’s robustness against artifacts and inter-subject variability. A transformer encoder then models temporal relationships across all epochs,  with attention mechanisms identifying the most informative time segments and EEG channels. This enhances interpretability and provides clinically meaningful relevance maps. Finally, a prediction module estimates either the likelihood of epilepsy progression within a defined time horizon or a risk score for survival analysis. Joint training ensures that both generative robustness and discriminative accuracy contribute to reliable prognosis.

Keywords

EEG; Hybrid diffusion-transformer model; Epilepsy progression; Feature extraction; EEG rhythm-based prognosis of epilepsy

1. Addison, P. S. (2017). The illustrated wavelet transform handbook: introductory theory and applications in science, engineering, medicine and finance. CRC press.
2. Alalayah, K. M., Senan, E. M., Atlam, H. F., Ahmed, I. A. & Shatnawi, H. S. (2023). Effective early detection of epileptic seizures through eeg signals using classification algorithms based on t-distributed stochastic neighbor embedding and K-means. Diagnostics, 13(11), 1957. Retrieved from https://doi.org/10.3390/diagnostics13111957
3. Berrich, Y. & Guennoun, Z. (2025). EEG-based epilepsy detection using CNN-SVM and DNN-SVM with feature dimensionality reduction by PCA. Scientific Reports, 15(1), 14313.
4. Gallotto, S. & Seeck, M. (2022). EEG biomarker candidates for the identification of epilepsy. Clinical Neurophysiology Practice, 8, 32-41.
5. Hilden, J. (1991). The area under the ROC curve and its competitors. Medical Decision Making, 11(2), 95-101.
6. Ho, J., Jain, A. & Abbeel, P. (2020). Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33, 6840-6851.
7. Huang, X., Li, C., Liu, A., Qian, R., & Chen, X. (2025). EEGDfus: A conditional diffusion model for fine-grained EEG denoising. IEEE Journal of Biomedical and Health Informatics, 29(4), 2557–2569. Retrieved from https://doi.org/10.1109/JBHI.2024.3504716
8. Lehnertz, K. (2008). Epilepsy and nonlinear dynamics. Journal of biological physics, 34(3), 253-266.
9. Ma, Y., Liu, C., Ma, M. S., Yang, Y., Truong, N. D., Kothur, K., Nikpour, A. & Kavehei, O. (2023). TSD: Transformers for seizure detection. Retrieved from https://doi.org/10.1101 /2023.01.24.525308
10. Montoya, X., Díaz, F., Félix, J., Paucar, J., Ferrer, J., & Fonseca, P. (2023). Seizure detection by analyzing the number of channels selected by cross-correlation using TUH EEG seizure corpus. In Proceedings of the 18th International Symposium on Medical Information Processing and Analysis (Vol. 12567, Article 1256710).
11. Niedermeyer, E. & da Silva, F. L. (Eds.). (2005). Electroencephalography: basic principles, clinical applications, and related fields. Lippincott Williams & Wilkins.
12. Parija, S., Raiguru, P. & Sahani, M. (2024). Automatic seizure detection of EEG signals using deep belief network with robust LMD-based features. 2024 3^rd Odisha International Conference on Electrical Power Engineering, Communication and Computing Technology (ODICON), 1-5.
13. Pfeffer, M. A., Wong, J. K. W., & Ling, S. H. (2025). Trends and limitations in transformer-based BCI research. Applied Sciences, 15(20), 11150.
14. Rasheed, K., Qayyum, A., Qadir, J., Sivathamboo, S., Kwan, P., Kuhlmann, L., O’Brien, T. & Razi, A. (2020). Machine learning for predicting epileptic seizures using EEG signals: A review. IEEE Reviews in Biomedical Engineering, 14, 139-155.
15. 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.
16. Sanei, S. & Chambers, J. A. (2013). EEG signal processing. John Wiley & Sons.
17. Snoek, J., Larochelle, H. & Adams, R. P. (2012). Practical Bayesian optimization of machine learning algorithms. Advances in Neural Information Processing Systems, 25, 2951–2959.
18. Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S. & Poole, B. (2020). Score-based generative modeling through stochastic differential equations. Retrieved from https://doi.org/10.48550/arXiv.2011.13456
19. Subha, D., Joseph, P., Acharya U., R. & Lim, C. (2010). EEG signal analysis: A survey. Journal of Medical Systems, 34, 195-212.
20. Thodoroff, P., Pineau, J. & Lim, A. (2016, December). Learning robust features using deep learning for automatic seizure detection. In Machine learning for healthcare conference
21. Torma, S. & Szegletes, L. (2025). Generative modeling and augmentation of EEG signals using improved diffusion probabilistic models. Journal of Neural Engineering, 22(1), 016001.
22. Tran, L. V., Tran, H. M., Le, T. M., Huynh, T. T., Tran, H. T. & Dao, S. V. (2022). Application of machine learning in epileptic seizure detection. Diagnostics, 12(11), 2879.
23. Uppalapati, K., Yimenicioglu, B., Abdulkareem, S., Uppalapati, B., Kamath, V., Eftekhari, A. & Ayyappan, P. (2025). A two-feature quantitative EEG index of pediatric epilepsy severity: External pre-validation on CHB-MIT and roadmap to Dravet cohorts. Retrieved from https://doi.org/10.48550/arXiv.2510.13815
24. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł. & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30, 5998–6008.
25. Yang, C., Wang, T. & Yan, X. (2023). DDMT: Denoising diffusion mask transformer models for multivariate time series anomaly detection. Retrieved from https://doi.org/10.4 8550/arXiv.2310.08800
26. Zhang, H. & Li, H. (2025). Transformer-based EEG decoding: A Survey. Retrieved from https://doi.org/10.48550/arXiv.2507.02320