Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation
<italic>Goal:</italic> Building a DL model that can be trained on small EEG training set of a single subject presents an interesting challenge that this work is trying to address. In particular, this study is trying to avoid the need for long EEG data collection sessions, and without com...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Open Journal of Engineering in Medicine and Biology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9962758/ |
| _version_ | 1797893826431418368 |
|---|---|
| author | Haider Alwasiti Mohd Zuki Yusoff |
| author_facet | Haider Alwasiti Mohd Zuki Yusoff |
| author_sort | Haider Alwasiti |
| collection | DOAJ |
| description | <italic>Goal:</italic> Building a DL model that can be trained on small EEG training set of a single subject presents an interesting challenge that this work is trying to address. In particular, this study is trying to avoid the need for long EEG data collection sessions, and without combining multiple subjects training datasets, which has a detrimental effect on the classification performance due to the inter-individual variability among subjects. <italic>Methods:</italic> A customized Convolutional Neural Network with mixup augmentation was trained with <inline-formula><tex-math notation="LaTeX">$\scriptstyle \mathtt {\sim }$</tex-math></inline-formula>120 EEG trials for only one subject per model. <italic>Results:</italic> Modified ResNet18 and DenseNet121 models with mixup augmentation achieved 0.920 (95% Confidence Interval: 0.908, 0.933) and 0.933 (95% Confidence Interval: 0.922, 0.945) classification accuracy, respectively. <italic>Conclusions:</italic> We show that the designed classifiers resulted in a higher classification performance in comparison to other DL classifiers of previous studies on the same dataset, despite the limited training dataset used in this work. |
| first_indexed | 2024-04-10T06:59:16Z |
| format | Article |
| id | doaj.art-2d5d2fc291c9496eb47bcd863db8eda1 |
| institution | Directory Open Access Journal |
| issn | 2644-1276 |
| language | English |
| last_indexed | 2024-04-10T06:59:16Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Engineering in Medicine and Biology |
| spelling | doaj.art-2d5d2fc291c9496eb47bcd863db8eda12023-02-28T00:01:11ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762022-01-01317117710.1109/OJEMB.2022.32201509962758Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup AugmentationHaider Alwasiti0https://orcid.org/0000-0002-2978-5492Mohd Zuki Yusoff1https://orcid.org/0000-0001-9306-6655Helsinki Lab of Interdisciplinary Conservation Science, University of Helsinki, Helsinki, FinlandCentre for Intelligent Signal and Imaging Research (CISIR), Department of Electrical and Electronic Engineering, Universiti Teknologi PETRONAS, Seri Iskandar, Perak, Malaysia<italic>Goal:</italic> Building a DL model that can be trained on small EEG training set of a single subject presents an interesting challenge that this work is trying to address. In particular, this study is trying to avoid the need for long EEG data collection sessions, and without combining multiple subjects training datasets, which has a detrimental effect on the classification performance due to the inter-individual variability among subjects. <italic>Methods:</italic> A customized Convolutional Neural Network with mixup augmentation was trained with <inline-formula><tex-math notation="LaTeX">$\scriptstyle \mathtt {\sim }$</tex-math></inline-formula>120 EEG trials for only one subject per model. <italic>Results:</italic> Modified ResNet18 and DenseNet121 models with mixup augmentation achieved 0.920 (95% Confidence Interval: 0.908, 0.933) and 0.933 (95% Confidence Interval: 0.922, 0.945) classification accuracy, respectively. <italic>Conclusions:</italic> We show that the designed classifiers resulted in a higher classification performance in comparison to other DL classifiers of previous studies on the same dataset, despite the limited training dataset used in this work.https://ieeexplore.ieee.org/document/9962758/EEGdeep learningBCIstockwell transform |
| spellingShingle | Haider Alwasiti Mohd Zuki Yusoff Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation IEEE Open Journal of Engineering in Medicine and Biology EEG deep learning BCI stockwell transform |
| title | Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation |
| title_full | Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation |
| title_fullStr | Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation |
| title_full_unstemmed | Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation |
| title_short | Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation |
| title_sort | motor imagery classification for brain computer interface using deep convolutional neural networks and mixup augmentation |
| topic | EEG deep learning BCI stockwell transform |
| url | https://ieeexplore.ieee.org/document/9962758/ |
| work_keys_str_mv | AT haideralwasiti motorimageryclassificationforbraincomputerinterfaceusingdeepconvolutionalneuralnetworksandmixupaugmentation AT mohdzukiyusoff motorimageryclassificationforbraincomputerinterfaceusingdeepconvolutionalneuralnetworksandmixupaugmentation |