A novel transformer for attention decoding using EEG
Electroencephalography (EEG) attention classification plays a crucial role in brain-computer interface (BCI) applications. This paper introduces EEG-PatchFormer, a novel deep learning model leveraging transformers to achieve superior EEG attention decoding. We posit that transformers’ strength in ca...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/175057 |
| _version_ | 1826121031114293248 |
|---|---|
| author | Lee, Joon Hei |
| author2 | Guan Cuntai |
| author_facet | Guan Cuntai Lee, Joon Hei |
| author_sort | Lee, Joon Hei |
| collection | NTU |
| description | Electroencephalography (EEG) attention classification plays a crucial role in brain-computer interface (BCI) applications. This paper introduces EEG-PatchFormer, a novel deep learning model leveraging transformers to achieve superior EEG attention decoding. We posit that transformers’ strength in capturing long-range temporal dependencies, coupled with their recent success on spatial data, makes them ideally suited for processing EEG signals.
We begin by outlining a pilot study investigating the impact of various patching strategies on the classification accuracy of a transformer-based network. This study revealed significant performance variations across patching methods, emphasising the importance of optimal patching for model efficacy.
We then showcase the proposed EEG-PatchFormer architecture. Key modules include a temporal convolutional neural network (CNN), a pointwise convolutional layer, and separate patching modules to handle global and local spatial features, as well as temporal features. The model then features a transformer module, and culminates in a fully-connected classifier.
Finally, EEG-PatchFormer’s performance across various evaluation experiments is discussed. Extensive evaluation on a publicly available cognitive attention dataset demonstrated that EEG-PatchFormer surpasses existing state-of-the-art benchmarks in terms of mean classification accuracy, area under the ROC curve (AUC), and macro-F1 score. Hyperparameter tuning and ablation studies were carried out to further optimise, and understand the contribution of, individual components.
Overall, this project establishes EEG-PatchFormer as a state-of-the-art model for EEG attention decoding, with promising applications for BCI. |
| first_indexed | 2024-10-01T05:26:06Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/175057 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T05:26:06Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1750572024-04-19T15:45:41Z A novel transformer for attention decoding using EEG Lee, Joon Hei Guan Cuntai School of Computer Science and Engineering CTGuan@ntu.edu.sg Computer and Information Science EEG Deep learning Attention Brain-computer interface Electroencephalography (EEG) attention classification plays a crucial role in brain-computer interface (BCI) applications. This paper introduces EEG-PatchFormer, a novel deep learning model leveraging transformers to achieve superior EEG attention decoding. We posit that transformers’ strength in capturing long-range temporal dependencies, coupled with their recent success on spatial data, makes them ideally suited for processing EEG signals. We begin by outlining a pilot study investigating the impact of various patching strategies on the classification accuracy of a transformer-based network. This study revealed significant performance variations across patching methods, emphasising the importance of optimal patching for model efficacy. We then showcase the proposed EEG-PatchFormer architecture. Key modules include a temporal convolutional neural network (CNN), a pointwise convolutional layer, and separate patching modules to handle global and local spatial features, as well as temporal features. The model then features a transformer module, and culminates in a fully-connected classifier. Finally, EEG-PatchFormer’s performance across various evaluation experiments is discussed. Extensive evaluation on a publicly available cognitive attention dataset demonstrated that EEG-PatchFormer surpasses existing state-of-the-art benchmarks in terms of mean classification accuracy, area under the ROC curve (AUC), and macro-F1 score. Hyperparameter tuning and ablation studies were carried out to further optimise, and understand the contribution of, individual components. Overall, this project establishes EEG-PatchFormer as a state-of-the-art model for EEG attention decoding, with promising applications for BCI. Bachelor's degree 2024-04-19T01:36:34Z 2024-04-19T01:36:34Z 2024 Final Year Project (FYP) Lee, J. H. (2024). A novel transformer for attention decoding using EEG. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/175057 https://hdl.handle.net/10356/175057 en SCSE23-0162 application/pdf Nanyang Technological University |
| spellingShingle | Computer and Information Science EEG Deep learning Attention Brain-computer interface Lee, Joon Hei A novel transformer for attention decoding using EEG |
| title | A novel transformer for attention decoding using EEG |
| title_full | A novel transformer for attention decoding using EEG |
| title_fullStr | A novel transformer for attention decoding using EEG |
| title_full_unstemmed | A novel transformer for attention decoding using EEG |
| title_short | A novel transformer for attention decoding using EEG |
| title_sort | novel transformer for attention decoding using eeg |
| topic | Computer and Information Science EEG Deep learning Attention Brain-computer interface |
| url | https://hdl.handle.net/10356/175057 |
| work_keys_str_mv | AT leejoonhei anoveltransformerforattentiondecodingusingeeg AT leejoonhei noveltransformerforattentiondecodingusingeeg |