---

Journal of Applied Nonlinear Dynamics 14(4) (2025) 859--874 | DOI:10.5890/JAND.2025.12.008

Wen Yu, Tianfeng Wang, Yeyin Xu

School of Aerospace Engineering, Xi'an Jiaotong University, Xi'an, 710049, PR China

Download Full Text PDF

 

Abstract

Epilepsy is a common brain disease in the world, and its clinical diagnosis has always been essential in epilepsy treatment. However, the current medical practice mainly relies on the patient's clinical symptoms and the doctor's manual judgment based on electroencephalogram(EEG), which has significant shortcomings in efficiency and accuracy. With the development of machine learning, artificial intelligence is used for the diagnosis of epileptic seizures in practice. In this paper, we established two different input features: only power spectral density (PSD) and recurrence rate (RR) combined with PSD. Such recurrence rate is achieved through the phase state reconstruction of the time-delayed signals. Considering the recursive characteristics, the accuracy for diagnosing epileptic seizures can reach to 99\%. We also compared the performance of the two types of input data and discussed the scale of input data on the results. The influence of time-delay parameters in the recurrence rate is studied. The result shows that the diagnosis through power spectral density with recursive rate as dataset is better than that of judging only by power spectral density. By considering the recurrence rate, the intelligent diagnosis accuracy of the neural network improves by varying degrees, ranging from 0.3\% to 1.5\%, depending on the size of the input data. The specificity of the neural network improves by varying degrees, ranging from 0.2\% to 3.4\%, depending on the size of the input data. Simultaneously, the sensitivity values also have risen, ranging from 0.1\% to 1.7\%. With the increase of data scale, the convergence speed of the neural network in earlier iteration stage will be accelerated. In Fig.10 (c) and (d), considering both RR and PSD results in achieving an accuracy of 98\% after only 32 iterations, which is comparable to the accuracy achieved after 186 iterations when only considering PSD. In Fig.11 (a) and (b), the curve of 400 PSD with or without RR is the fastest descending curve in their series, which reflects that the convergence speed accelerates as dataset size increases. However, when the data volume gets larger, the CNN network needs more iteration epochs to achieve stable. That's why there is more fluctuation in curve of 400 PSD with or without RR. The research provides a good perspective for clinical diagnosis of epilepsy seizures.

References

1. [1]	Takami, Y., Satake, E., and Ban, H. (2015), Risk of seizure recurrence after a first unprovoked seizure in childhood, No to Hattatsu Brain and Development, 47(6), 427-432.

2. [2]	Beniczky, S. and Ryvlin, P. (2018), Standards for testing and clinical validation of seizure detection devices, Epilepsia, 59, 9-13.

3. [3]	Elger, C.E. and Hoppe, C. (2018), Diagnostic challenges in epilepsy: seizure under-reporting and seizure detection, The Lancet Neurology, 17(3), 279-288.

4. [4]	Gotman, J. (1982), Automatic recognition of epileptic seizures in the EEG, Electroencephalography and Clinical Neurophysiology, 54(5), 530-540.

5. [5]	Murro, A.M., King, D.W., Smith, J.R., Gallagher, B.B., Flanigin, H.F., and Meador, K. (1991), Computerized seizure detection of complex partial seizures, Electroencephalography and Clinical Neurophysiology, 79(4), 330-333.

6. [6]	Qu, H. and Gotman, J. (1997), A patient-specific algorithm for the detection of seizure onset in long-term EEG monitoring: possible use as a warning device, IEEE Transactions on Biomedical Engineering, 44(2), 115-122.

7. [7]	Litjens, G., Kooi, T., Bejnordi, B.E., Setio, A.A.A., Ciompi, F., Ghafoorian, M., Van Der Laak, J.A., Van Ginneken, B., and Sánchez, C.I. (2017), A survey on deep learning in medical image analysis, Medical Image Analysis, 42, 60-88.

8. [8]	Xun, G., Jia, X., and Zhang, A. (2016), Detecting epileptic seizures with electroencephalogram via a context-learning model, BMC Medical Informatics and Decision Making, 16, 97-109.

9. [9]	Subasi, A. and Ercelebi, E. (2005), Classification of EEG signals using neural network and logistic regression, Computer Methods and Programs in Biomedicine, 78(2), 87-99.

10. [10]	Mirowski, P.W., LeCun, Y., Madhavan, D., and Kuzniecky, R. (2008), Comparing SVM and convolutional networks for epileptic seizure prediction from intracranial EEG, in 2008 IEEE Workshop on Machine Learning for Signal Processing, (pp. 244-249). IEEE.

11. [11]	Acharya, U.R., Oh, S.L., Hagiwara, Y., Tan, J.H., and Adeli, H. (2018), Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals, Computers in Biology and Medicine, 100, 270-278.

12. [12]	Shoeb, A.H. (2009), Application of machine learning to epileptic seizure onset detection and treatment, Doctoral dissertation, Massachusetts Institute of Technology.

13. [13]	Zhang, Z., Chen, Z., Zhou, Y., Du, S., Zhang, Y., Mei, T., and Tian, X. (2014), Construction of rules for seizure prediction based on approximate entropy, Clinical Neurophysiology, 125(10), 1959-1966.

14. [14]	Youngworth, R.N., Gallagher, B.B., and Stamper, B.L. (2005), An overview of power spectral density (PSD) calculations, Optical Manufacturing and Testing VI, 5869, 206-216.

15. [15]	Packard, N.H., Crutchfield, J.P., Farmer, J.D., and Shaw, R.S. (1980), Geometry from a time series, Physical Review Letters, 45(9), 712.

16. [16]	Noakes, L. (1991), The Takens embedding theorem, International Journal of Bifurcation and Chaos, 1(04), 867-872.

17. [17]	Marwan, N., Romano, M.C., Thiel, M., and Kurths, J. (2007), Recurrence plots for the analysis of complex systems, Physics Reports, 438(5-6), 237-329.

18. [18]	Fukushima, K. (1980), Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position, Biological Cybernetics, 36(4), 193-202.

19. [19]	Waibel, A. (1987), Phoneme recognition using time-delay neural networks, Meeting of the Institute of Electrical, Information and Communication Engineers (IEICE), Tokyo, Japan.

20. [20]	Suzuki, K. (2017), Overview of deep learning in medical imaging, Radiological Physics and Technology, 10(3), 257-273.

21. [21]	Johansen, A.R., Jin, J., Maszczyk, T., Dauwels, J., Cash, S.S., and Westover, M.B. (2016, March), Epileptiform spike detection via convolutional neural networks, in 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 754-758). IEEE.

22. [22]	Aliyu, I. and Lim, C.G. (2023), Selection of optimal wavelet features for epileptic EEG signal classification with LSTM, Neural Computing and Applications, 1-21.

23. [23]	Zeng, J., Tan, X.D. and Chang'an, A.Z. (2021), Automatic detection of epileptic seizure events using the time-frequency features and machine learning, Biomedical Signal Processing and Control, 69, 102916.

24. [24]	Nasiri, S. and Clifford, G.D. (2021), Generalizable seizure detection model using generating transferable adversarial features, IEEE Signal Processing Letters, 28, 568-572.

25. [25]	Zeng, K., Ouyang, G., Chen, H., Gu, Y., Liu, X., and Li, X. (2018), Characterizing dynamics of absence seizure EEG with spatial-temporal permutation entropy, Neurocomputing, 275, 577-585.

26. [26]	Solaija, M.S.J., Saleem, S., Khurshid, K., Hassan, S.A., and Kamboh, A.M. (2018), Dynamic mode decomposition based epileptic seizure detection from scalp EEG, IEEE Access, 6, 38683-38692.

27. [27]	Lu, X., Zhang, J., Huang, S., Lu, J., Ye, M., and Wang, M. (2021), Detection and classification of epileptic EEG signals by the methods of nonlinear dynamics, Chaos, Solitons \& Fractals, 151, 111032.

28. [28]

    Moravvej, S.V., Alizadehsani, R., Khanam, S., Sobhaninia, Z., Shoeibi, A., Khozeimeh, F., Sani, Z.A., Tan, R.S., Khosravi, A., Nahavandi, S., and Kadri, N.A. (2022), RLMD-PA: a reinforcement learning-based myocarditis diagnosis combined with a population-based algorithm for pretraining weights, Contrast Media $\&$ Molecular Imaging, 2022(1), 8733632.

29. [29]

    Shoeibi, A., Ghassemi, N., Khodatars, M., Moridian, P., Khosravi, A., Zare, A., Gorriz, J.M., Chale-Chale, A.H., Khadem, A., and Rajendra Acharya, U. (2023), Automatic diagnosis of schizophrenia and attention deficit hyperactivity disorder in rs-fMRI modality using convolutional autoencoder model and interval type-2 fuzzy regression, Cognitive Neurodynamics, 17(6), 1501-1523.