Analyzing Shallow and Deep CNNs with Gold Standard and Random EEG Segment Selections for Coma Prognosis

• Gustavo B. S. Oliveira UFU
• Paulo D. S. Souza UFU
• Tiago S. Silva UFU
• João B. D. Filho UFU
• Murillo G. Carneiro UFU

DOI: https://doi.org/10.5753/sbcas.2025.7669

Resumo

This study focuses on deep learning models for predicting coma outcomes using electroencephalogram (EEG) data, exploring convolutional neural networks (CNNs), particularly Shallow and Deep ConvNets, based on Filter Bank Common Spatial Patterns. A dataset of 121 EEG samples (42 favorable, 79 unfavorable) was analyzed. EEG segments were selected using two strategies and frequencies. Models were trained with 10-fold cross-validation and FTSurrogate for class balance. Shallow ConvNet showed stable performance across frequencies, while Deep ConvNet excelled at 200Hz. Simple segment selection and sampling frequency methods improved CNN performance. The findings offer insights for future research and potential clinical applications.

Referências

Aellen, F. M., Alnes, S. L., Loosli, F., Rossetti, A. O., Zubler, F., De Lucia, M., and Tzovara, A. (2023). Auditory stimulation and deep learning predict awakening from coma after cardiac arrest. Brain, 146(2):778–788.

Akiba, T., Sano, S., Yanase, T., Ohta, T., and Koyama, M. (2019). Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pages 2623–2631.

Al-Hussaini, I. and Mitchell, C. S. (2023). Seizft: interpretable machine learning for seizure detection using wearables. Bioengineering, 10(8):918.

Baldo Júnior, S., Carneiro, M. G., Destro-Filho, J.-B., Zhao, L., and Tinos, R. (2023). Classification of coma etiology using convolutional neural networks and long-short term memory networks. In 2023 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE.

Bissaro, L. Z. (2021). Aprendizado de padrões eeg para prognóstico precoce de pacientes em coma usando redes echo state e redes neurais convolucionais. Master’s thesis, Universidade Federal de Uberlândia, Uberlândia.

Bissaro, L. Z., Junior, O. O. N., Destro Filho, J. B., Jin, Y., and Carneiro, M. G. (2023). Towards the prognosis of patients in coma using echo state networks for eeg analysis. Procedia Computer Science, 222:509–518.

Carneiro, M. G., Ramos, C. D., Destro-Filho, J.-B., Zhu, Y.-t., Ji, D., and Zhao, L. (2023). High-level classification for eeg analysis. In 2023 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE.

Caza-Szoka, M. and Massicotte, D. (2022). Windowing compensation in fourier based surrogate analysis and application to eeg signal classification. IEEE Transactions on Instrumentation and Measurement, 71:1–11.

Coelli, S., Calcagno, A., Cassani, C. M., Temporiti, F., Reali, P., Gatti, R., Galli, M., and Bianchi, A. M. (2024). Selecting methods for a modular eeg pre-processing pipeline: An objective comparison. Biomedical Signal Processing and Control, 90:105830.

Costa, P. G. d. et al. (2022). Base de dados em eletroencefalografia (eeg).

de Paiva, M. R., Gobbo Jr, M., Rufim, E. S., Campos, M., and Destro Filho, J. B. (2018). Avaliação visual dos padrões eletroencefalográficos de pacientes clinicamente em coma. Brazilian Journal of Health Review, 1(2):447–455.

Delorme, A. (2023). Eeg is better left alone. Scientific reports, 13(1):2372.

Gramfort, A., Luessi, M., Larson, E., Engemann, D. A., Strohmeier, D., Brodbeck, C., Goj, R., Jas, M., Brooks, T., Parkkonen, L., et al. (2013). Meg and eeg data analysis with mne-python. Frontiers in neuroscience, 7:70133.

Hossain, K. M., Islam, M. A., Hossain, S., Nijholt, A., and Ahad, M. A. R. (2023). Status of deep learning for eeg-based brain–computer interface applications. Frontiers in computational neuroscience, 16:1006763.

İnce, R., Adanır, S. S., and Sevmez, F. (2021). The inventor of electroencephalography (eeg): Hans berger (1873–1941). Child’s Nervous System, 37:2723–2724.

Iwama, S., Morishige, M., Kodama, M., Takahashi, Y., Hirose, R., and Ushiba, J. (2023). High-density scalp electroencephalogram dataset during sensorimotor rhythm-based brain-computer interfacing. Scientific Data, 10(1):385.

LeCun, Y., Bengio, Y., and Hinton, G. (2015). Deep learning. nature, 521(7553):436–444.

Mathew, A., Amudha, P., and Sivakumari, S. (2021). Deep learning techniques: An overview. Advanced Machine Learning Technologies and Applications, page 599.

Ramos, C. D. et al. (2022). Aprendizado de máquina como ferramenta para o prognóstico de pacientes em coma usando sinais eletroencefalográficos no espectro de 1 a 100 hz.

Saeidi, M., Karwowski, W., Farahani, F. V., Fiok, K., Taiar, R., Hancock, P. A., and Al-Juaid, A. (2021). Neural decoding of eeg signals with machine learning: A systematic review. Brain Sciences, 11(11):1525.

Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J., Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F., Burgard, W., and Ball, T. (2017). Deep learning with convolutional neural networks for eeg decoding and visualization. Human brain mapping, 38(11):5391–5420.

Xu, M., Ouyang, Y., and Yuan, Z. (2023). Deep learning aided neuroimaging and brain regulation. Sensors, 23(11):4993.

Yang, L. and Shami, A. (2020). On hyperparameter optimization of machine learning algorithms: Theory and practice. Neurocomputing, 415:295–316.

Yu, T. and Zhu, H. (2020). Hyper-parameter optimization: A review of algorithms and applications. arXiv preprint arXiv:2003.05689.