Rahma Maalej; Abir Hadriche; Mohamed Amine Ben Msarra; Nawel Jmail

A Feature-Enhanced Hybrid CNN-BiLSTM Framework for Multi-Label Classification of Pathological High-Frequency Oscillations in Intracranial EEG Signals

PDF (1358KB), PP.1-16

Views: 0 Downloads: 0

Author(s)

Rahma Maalej ^1,* Abir Hadriche ² Mohamed Amine Ben Msarra ¹ Nawel Jmail ³

1. University of Sfax, Miracl Lab, Department of STIC, National School of Electronics and Telecommunications of Sfax, Tunisia

2. University of Sfax, Research Groups in Intelligent Machines, Regim Lab, Digital Research Center of Sfax, CRNS, Enis, High institute of Music, Sfax, Tunisia

3. University of Sfax, Miracl Lab, Digital Research Center of Sfax, CRNS, Higher Business School, Sfax, Tunisia

* Corresponding author.

DOI: https://doi.org/10.5815/ijigsp.2026.03.01

Received: 5 Jan. 2026 / Revised: 19 Feb. 2026 / Accepted: 10 Mar. 2026 / Published: 8 Jun. 2026

Index Terms

High-Frequency Oscillations, Intracranial Electroencephalography, Multi-Label Classification, Hybrid Deep Learning Architecture, Nonlinear Signal Analysis, Epileptogenic Zone Localization

Abstract

Interictal high-frequency oscillations, including ripples in the frequency range of 80-250 hertz and fast ripples between 250-500 hertz, are increasingly recognized as reliable electrophysiological biomarkers for delineating the epileptogenic zone in patients with drug-resistant epilepsy. However, their routine clinical exploitation remains limited due to pronounced morphological variability, low signal-to-noise ratios, and the difficulty of identifying overlapping events in which ripples and fast ripples occur simultaneously.
This paper presents an automated deep learning framework designed for the multi-label classification of pathological high-frequency oscillations in intracranial electroencephalographic signals. The proposed approach integrates advanced nonlinear statistical descriptors, including entropy- and complexity-based measures, in order to enhance the discriminative representation of the signals. These features are processed using a hybrid deep learning architecture that combines convolutional neural networks for local morphological feature extraction with bidirectional long short-term memory networks to capture long-range temporal dependencies in non-stationary neural signals.
The proposed framework was evaluated using the publicly available multi-patient intracranial electroencephalography dataset provided by the Collaborative Research in Computational Neuroscience initiative. Experimental results demonstrate a classification accuracy of 98.3 %, along with high precision and balanced performance across all pathological classes. These findings indicate that the proposed method offers a robust and objective solution for the automated identification of high-frequency oscillations, with strong potential for improving presurgical evaluation and decision-making in epilepsy surgery.

Cite This Paper

Rahma Maalej, Abir Hadriche, Mohamed Amine Ben Msarra, Nawel Jmail, "A Feature-Enhanced Hybrid CNN-BiLSTM Framework for Multi-Label Classification of Pathological High-Frequency Oscillations in Intracranial EEG Signals", International Journal of Image, Graphics and Signal Processing(IJIGSP), Vol.18, No.3, pp. 1-16, 2026. DOI:10.5815/ijigsp.2026.03.01

Reference

[1]World Health Organization, “Epilepsy,” World Health Organization (WHO) Fact Sheets, 2019. [Online]. Available: https://www.who.int/health-topics/epilepsy. [Accessed: Sep. 23, 2025].
[2]H. Suzuki et al., “Identification of drug-resistant laterality in bilateral temporal lobe epilepsy,” Cureus, vol. 17, no. 5, 2025.
[3]W. U. R. Qamar, M.-H. Lee, and B. Abibullaev, “Deep learning in intracranial eeg for seizure detection: advances, challenges, and clinical applications,” Front. Neurosci., vol. 19, p. 1677898, 2025.
[4]A. Daida et al., “Ai-based localization of the epileptogenic zone using intracranial eeg,” Epilepsia Open, 2025.
[5]J. Cheng et al., “Localization of epileptogenic zone from seeg: Combination of high-frequency energy and synchronous connection in epileptic network analysis,” Biomed. Signal Process. Control, vol. 100, p. 107056, 2025.
[6]C. Du et al., “Automated detection of the epileptogenic zone in stereoelectroencephalography for drug-resistant epilepsy using multi-epileptogenic biomarker machine learning,” Epilepsy Res., p. 107710, 2025.
[7]ELBehy, A. Hadriche, R. Maalej, and N. Jmail, “Comparison between different source localization and connectivity metrics of spiky and oscillatory meg activities,” Int. J. Electr. Comput. Eng. Syst., vol. 15, no. 10, pp. 875–884, 2024.
[8]A. Raghu et al., “Machine learning-based localization of the epileptogenic zone using high-frequency oscillations from seeg: A real-world approach,” J. Clin. Neurosci., vol. 135, p. 111177, 2025.
[9]N. Jmail, R. Jarray, A. Hadriche, T. Frikha, and C. Benar, “Separation between spikes and oscillation by stationary wavelet transform implemented on an embedded architecture,” J. Neurol. Sci., vol. 381, p. 542, 2017.
[10]A. Aymen, S. E. Khediri, A. Thaljaoui, M. Miladi, and A. Kachouri, “Catalyzing eeg signal analysis: unveiling the potential of machine learning-enabled smart k nearest neighbor outlier detection,” Int. J. Inf. Technol., vol. 17, no. 9, pp. 5613–5624, 2025.
[11]Z. Gu, S. Yang, Z. Yu, and F. Liu, “Detection of high-frequency oscillations from intracranial eeg data with switching state space model,” in 2024 46th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (EMBC), 2024, pp. 1–4.
[12]A. Hadriche, N. Jmail, A. Kachouri, and H. Ghariani, “The detection of evoked potential with variable latency and multiple trial using consensus matching pursuit,” in Proc. 1st Int. Conf. Adv. Technol. Signal Image Process. (ATSIP), Sfax, Tunisia, 2014, pp. 316–321.
[13]R. Maalej, A. Hadriche, and N. Jmail, “Artificial intelligence in epilepsy: A narrative review of automated high-frequency oscillation detection and seizure recognition (2022–2025),” Epilepsy & Seizure, vol. 18, p. A000175, 2026.
[14]Z. Sadek, A. Hadriche, R. Maalej, and N. Jmail, “Multi-classification of high-frequency oscillations using ieeg signals and deep learning models,” arXiv preprint arXiv:2412.17145, 2024.
[15]G. Yi, J. Song, W. Zhang, J. Wang, S. Li, and L. Cai, “Cnn with double-side weighted visibility graph for automated classification of high-frequency oscillations in epilepsy,” Biomed. Signal Process. Control, vol. 110, p. 108178, 2025.
[16]S. Aslan and H. Bingöl, “Epileptic seizure detection from eeg signals with recurrent neural networks based classification model,” J. Phys. Chem. Funct. Mater., vol. 7, no. 2, pp. 14–21, 2024.
[17]Z. Huang et al., “Eeg detection and recognition model for epilepsy based on dual attention mechanism,” Sci. Rep., vol. 15, no. 1, p. 9404, 2025.
[18]S. Chaibi and A. Kachouri, “Toward reliable models for distinguishing epileptic high-frequency oscillations (hfos) from non-hfo events using lstm and pre-trained owl-vit vision–language framework,” AI, vol. 6, no. 9, p. 230, 2025.
[19]J. Su et al., “A model for epileptic eeg detection and recognition based on multi-attention mechanism and spatiotemporal,” Sci. Rep., vol. 15, no. 1, p. 31993, 2025.
[20]V. Adusumilli and M. M. Bee, “Epileptic seizure detection and classification of eeg signal using k-nearest neighbors (knn) compared with anfis-adaptive network-based fuzzy inference system,” AIP Conf. Proc., vol. 2816, p. 030001, 2024.
[21]S. Padmakala, “Epileptic seizure detection in rodent IEEG using machine learning: A comparative study of random forest, svm, and gradient boosting models,” in 2025 9th Int. Conf. Inventive Syst. Control (ICISC), 2025, pp. 1164–1171.
[22]F. Krikid et al., “Multi-classification of high-frequency oscillations in intracranial eeg signals based on cnn and data augmentation,” Signal Image Video Process., vol. 18, no. 2, pp. 1099–1109, 2024.
[23]L. K. Kumari et al., “Advanced epileptic seizure recognition with a hybrid CNN–Bi-LSTM model on eeg signals,” in 2025 IEEE 1st Int. Conf. Smart Sustainable Dev. Elect. Eng. (SSDEE), 2025, pp. 1–5.
[24]W. Zhao et al., “Residual and bidirectional lstm for epileptic seizure detection,” Front. Comput. Neurosci., vol. 18, p. 1415967, 2024.
[25]T. Fedele et al., “High frequency oscillations detected in the intracranial eeg of epilepsy patients during interictal sleep, patients’ electrode location and outcome of epilepsy surgery,” CRCNS.org, 2017.
[26]S. Burnos et al., “The morphology of high frequency oscillations (hfo) does not improve delineating the epileptogenic zone,” Clin. Neurophysiol., vol. 127, no. 4, pp. 2140–2148, 2016.
[27]S. S. Sikarwar, A. K. Rana, and S. S. Sengar, “Entropy-driven deep learning framework for epilepsy detection using electro encephalogram signals,” Neuroscience, vol. 577, pp. 12–24, 2025.
[28]W. Chen et al., “Unsupervised detection of high-frequency oscillations in intracranial electroencephalogram: promoting a valuable automated diagnostic tool for epilepsy,” Front. Neurol., vol. 16, p. 1455613, 2025.
[29]S. Lu et al., “Leveraging channel coherence in long-term ieeg data for seizure prediction,” IEEE J. Biomed. Health Inform., 2025.
[30]R. Maalej, A. Hadriche, and N. Jmail, “A simulation-based cnn–transformer hybrid model for the classification of high-frequency oscillations in intracranial eeg signals,” Epilepsy & Seizure, vol. 17, p. A000170, 2025.

International Journal of Image, Graphics and Signal Processing (IJIGSP)