Poster No:
1138
Submission Type:
Abstract Submission
Authors:
arian yavari1
Institutions:
1bishop's university, sherbrooke, Quebec
First Author:
Introduction:
Electroencephalography (EEG) is a powerful tool for studying cognitive processes such as attention, perception, and emotion, particularly in response to visual stimuli. However, EEG faces challenges due to signal variability across subjects, frequency bands, and stimulus conditions. This study investigates how gamma-band performance, frequency-band interactions, and visual stimuli impact EEG classification accuracy. Specifically, we explore the gamma band (30–100 Hz) and the effects of varying contrast and randomization levels of visual stimuli on classification outcomes. We also examine the role of event-related spectral perturbations (ERSPs) in improving classification models. The integration of machine learning and deep learning techniques allows for more accurate, automated EEG classification, aiming to contribute to the development of personalized Brain-Computer Interface (BCI) applications.
Methods:
Data were collected from 29 participants, with a focus on 16 subjects exhibiting strong gamma responses (30–100 Hz). Visual stimuli varied in contrast (0%, 5%, 33%, 100%) and randomization levels (0%, 10%, 60%).
Preprocessing: EEG data were baseline-corrected, and only the most stable and common number of trials were selected to ensure consistency.
Feature Extraction:
• Deep Learning Models: Event-related spectral perturbations (ERSPs) and spectrograms (60 frequencies × 350 timepoints) were created as inputs for dense networks, Conv2D, LSTMs, GRUs, and RNNs.
• Machine Learning Models: Features were extracted on a per-subject and per-frequency basis for each classification task.
Classification Methodology:
• Deep Learning:
o Multi-Class: The analysis included all subjects, stimuli, frequencies, trials, and timepoints.
o Binary: For binary classification, subjects with strong gamma responses were used to distinguish between 100% and 5% contrast conditions. Models were trained for 30 epochs with a 70/30 train-validation split.
• Machine Learning:
o Per-Frequency: Multiple conditions were tested across subjects and trials, incorporating varying stimuli, frequencies, and timepoint windows.
o Per-Subject: Different combinations of stimuli, frequency bands, and time windows were tested across all trials.
Each dataset followed a 70/30 train-validation split.
Results:
Machine Learning Classification: Random Forest performed best, especially for subjects with strong gamma responses. Accuracy varied by subject, frequency band, and stimulus condition. Certain subjects consistently contributed to the model's predictability across conditions.
Deep Learning Classification: Dense networks performed best in both multi-class and binary tasks. Data augmentation and regularization techniques improved model performance. Subjects with strong gamma responses exhibited better accuracy in the 30-100 Hz range, especially during contrast stimuli.
Conclusions:
This study emphasizes the importance of gamma-band responses and spectrogram-based
approaches for EEG classification. The findings highlight the role of preprocessing, frequency band selection, and individual subject characteristics in achieving high classification accuracy. These insights could improve real-time EEG decoding for cognitive and clinical applications, paving the way for advanced BCI technologies.
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
EEG/MEG Modeling and Analysis 2
Multivariate Approaches
Keywords:
Cognition
Electroencephaolography (EEG)
Machine Learning
Multivariate
Neurotransmitter
Perception
Plasticity
1|2Indicates the priority used for review
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
I accept
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information.
Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
I am submitting this abstract as an original work to be reproduced. I am available to be the “source party” in an upcoming team and consent to have this work listed on the OSSIG website. I agree to be contacted by OSSIG regarding the challenge and may share data used in this abstract with another team.
Please indicate below if your study was a "resting state" or "task-activation” study.
Task-activation
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Healthy subjects
Was this research conducted in the United States?
No
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel?
NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Yes
Were any animal research approved by the relevant IACUC or other animal research panel?
NOTE: Any animal studies without IACUC approval will be automatically rejected.
No
Please indicate which methods were used in your research:
EEG/ERP
Provide references using APA citation style.
1.
Dauwels, J., Vialatte, F., & Cichocki, A. (2010). EEG neurofeedback: From brain training to cognitive state monitoring. IEEE Transactions on Biomedical Engineering, 57(12), 2924-2933. https://doi.org/10.1109/TBME.2010.2061105
2.
He, H., & Wu, D. (2018). Transfer learning for brain-computer interfaces: A Euclidean space data alignment approach. IEEE Transactions on Biomedical Engineering, 65(6), 1413-1420. https://doi.org/10.1109/TBME.2017.2725917
3.
Liu, Y., Sourina, O., & Nguyen, M. K. (2008). Real-time EEG signal processing and visualization for brain-computer interface. Proceedings of the International Conference on Computer Science and Information Technology, 473-478. https://doi.org/10.1109/ICCSIT.2008.121
4.
Murray, S. O., & Thut, G. (2011). The neural basis of brain oscillations. Current Opinion in Neurobiology, 21(6), 801-808. https://doi.org/10.1016/j.conb.2011.08.010
5.
Pfurtscheller, G., & Lopes da Silva, F. H. (1999). Event-related EEG/MEG synchronization and desynchronization: Basic principles. Clinical Neurophysiology, 110(11), 1842-1857. https://doi.org/10.1016/S1388-2457(99)00141-8
6.
Scherer, K. R., & Ekman, P. (2004). Handbook of Affective Sciences. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195115085.001.0001
7.
Shuai, Z., Zhuang, P., & Sun, S. (2020). EEG signal classification with deep neural networks for brain-computer interface applications. Journal of Neural Engineering, 17(4), 046011. https://doi.org/10.1088/1741-2552/ab8c93
8.
Zhang, D., & Li, X. (2019). Convolutional neural networks for EEG classification: A review. Neurocomputing, 358, 1-12. https://doi.org/10.1016/j.neucom.2019.03.076
9.
Ding, X., Zhao, J., Sang, T., & Lee, M. (2022). A deep learning method approach for sleep stage classification with EEG spectrogram. International Journal of Environmental Research and Public Health, 19(10), 6322. https://doi.org/10.3390/ijerph19106322
10.
Gao, Y., Xu, Z., & He, H. (2020). Deep learning-based electroencephalography analysis: A systematic review. arXiv:1901.05498.
11.
Zhang, D., & Li, X. (2019). Convolutional neural networks for EEG classification: A review. Neurocomputing, 358, 1-12. https://doi.org/10.1016/j.neucom.2019.03.076
No