Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models
IntroductionA challenge when applying an artificial intelligence (AI) deep learning (DL) approach to novel electroencephalography (EEG) data, is the DL architecture's lack of adaptability to changing numbers of EEG channels. That is, the number of channels cannot vary neither in the training da...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-01-01
|
| Series: | Frontiers in Neuroinformatics |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fninf.2023.1272791/full |
| _version_ | 1797339155680722944 |
|---|---|
| author | Thomas Tveitstøl Thomas Tveitstøl Mats Tveter Mats Tveter Ana S. Pérez T. Ana S. Pérez T. Christoffer Hatlestad-Hall Anis Yazidi Hugo L. Hammer Hugo L. Hammer Ira R. J. Hebold Haraldsen |
| author_facet | Thomas Tveitstøl Thomas Tveitstøl Mats Tveter Mats Tveter Ana S. Pérez T. Ana S. Pérez T. Christoffer Hatlestad-Hall Anis Yazidi Hugo L. Hammer Hugo L. Hammer Ira R. J. Hebold Haraldsen |
| author_sort | Thomas Tveitstøl |
| collection | DOAJ |
| description | IntroductionA challenge when applying an artificial intelligence (AI) deep learning (DL) approach to novel electroencephalography (EEG) data, is the DL architecture's lack of adaptability to changing numbers of EEG channels. That is, the number of channels cannot vary neither in the training data, nor upon deployment. Such highly specific hardware constraints put major limitations on the clinical usability and scalability of the DL models.MethodsIn this work, we propose a technique for handling such varied numbers of EEG channels by splitting the EEG montages into distinct regions and merge the channels within the same region to a region representation. The solution is termed Region Based Pooling (RBP). The procedure of splitting the montage into regions is performed repeatedly with different region configurations, to minimize potential loss of information. As RBP maps a varied number of EEG channels to a fixed number of region representations, both current and future DL architectures may apply RBP with ease. To demonstrate and evaluate the adequacy of RBP to handle a varied number of EEG channels, sex classification based solely on EEG was used as a test example. The DL models were trained on 129 channels, and tested on 32, 65, and 129-channels versions of the data using the same channel positions scheme. The baselines for comparison were zero-filling the missing channels and applying spherical spline interpolation. The performances were estimated using 5-fold cross validation.ResultsFor the 32-channel system version, the mean AUC values across the folds were: RBP (93.34%), spherical spline interpolation (93.36%), and zero-filling (76.82%). Similarly, on the 65-channel system version, the performances were: RBP (93.66%), spherical spline interpolation (93.50%), and zero-filling (85.58%). Finally, the 129-channel system version produced the following results: RBP (94.68%), spherical spline interpolation (93.86%), and zero-filling (91.92%).ConclusionIn conclusion, RBP obtained similar results to spherical spline interpolation, and superior results to zero-filling. We encourage further research and development of DL models in the cross-dataset setting, including the use of methods such as RBP and spherical spline interpolation to handle a varied number of EEG channels. |
| first_indexed | 2024-03-08T09:41:52Z |
| format | Article |
| id | doaj.art-8216ec604ec042b3b665f49ae8bb0927 |
| institution | Directory Open Access Journal |
| issn | 1662-5196 |
| language | English |
| last_indexed | 2024-03-08T09:41:52Z |
| publishDate | 2024-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroinformatics |
| spelling | doaj.art-8216ec604ec042b3b665f49ae8bb09272024-01-30T04:16:09ZengFrontiers Media S.A.Frontiers in Neuroinformatics1662-51962024-01-011710.3389/fninf.2023.12727911272791Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning modelsThomas Tveitstøl0Thomas Tveitstøl1Mats Tveter2Mats Tveter3Ana S. Pérez T.4Ana S. Pérez T.5Christoffer Hatlestad-Hall6Anis Yazidi7Hugo L. Hammer8Hugo L. Hammer9Ira R. J. Hebold Haraldsen10Department of Neurology, Oslo University Hospital, Oslo, NorwayInstitute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, NorwayDepartment of Neurology, Oslo University Hospital, Oslo, NorwayInstitute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, NorwayDepartment of Neurology, Oslo University Hospital, Oslo, NorwayInstitute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, NorwayDepartment of Neurology, Oslo University Hospital, Oslo, NorwayDepartment of Computer Science, Oslo Metropolitan University, Oslo, NorwayDepartment of Computer Science, Oslo Metropolitan University, Oslo, NorwayDepartment of Holistic Systems, SimulaMet, Oslo, NorwayDepartment of Neurology, Oslo University Hospital, Oslo, NorwayIntroductionA challenge when applying an artificial intelligence (AI) deep learning (DL) approach to novel electroencephalography (EEG) data, is the DL architecture's lack of adaptability to changing numbers of EEG channels. That is, the number of channels cannot vary neither in the training data, nor upon deployment. Such highly specific hardware constraints put major limitations on the clinical usability and scalability of the DL models.MethodsIn this work, we propose a technique for handling such varied numbers of EEG channels by splitting the EEG montages into distinct regions and merge the channels within the same region to a region representation. The solution is termed Region Based Pooling (RBP). The procedure of splitting the montage into regions is performed repeatedly with different region configurations, to minimize potential loss of information. As RBP maps a varied number of EEG channels to a fixed number of region representations, both current and future DL architectures may apply RBP with ease. To demonstrate and evaluate the adequacy of RBP to handle a varied number of EEG channels, sex classification based solely on EEG was used as a test example. The DL models were trained on 129 channels, and tested on 32, 65, and 129-channels versions of the data using the same channel positions scheme. The baselines for comparison were zero-filling the missing channels and applying spherical spline interpolation. The performances were estimated using 5-fold cross validation.ResultsFor the 32-channel system version, the mean AUC values across the folds were: RBP (93.34%), spherical spline interpolation (93.36%), and zero-filling (76.82%). Similarly, on the 65-channel system version, the performances were: RBP (93.66%), spherical spline interpolation (93.50%), and zero-filling (85.58%). Finally, the 129-channel system version produced the following results: RBP (94.68%), spherical spline interpolation (93.86%), and zero-filling (91.92%).ConclusionIn conclusion, RBP obtained similar results to spherical spline interpolation, and superior results to zero-filling. We encourage further research and development of DL models in the cross-dataset setting, including the use of methods such as RBP and spherical spline interpolation to handle a varied number of EEG channels.https://www.frontiersin.org/articles/10.3389/fninf.2023.1272791/fullEEGdeep learningmachine learningcross-datasetcross-channel systemconvolutional neural networks |
| spellingShingle | Thomas Tveitstøl Thomas Tveitstøl Mats Tveter Mats Tveter Ana S. Pérez T. Ana S. Pérez T. Christoffer Hatlestad-Hall Anis Yazidi Hugo L. Hammer Hugo L. Hammer Ira R. J. Hebold Haraldsen Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models Frontiers in Neuroinformatics EEG deep learning machine learning cross-dataset cross-channel system convolutional neural networks |
| title | Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models |
| title_full | Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models |
| title_fullStr | Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models |
| title_full_unstemmed | Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models |
| title_short | Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models |
| title_sort | introducing region based pooling for handling a varied number of eeg channels for deep learning models |
| topic | EEG deep learning machine learning cross-dataset cross-channel system convolutional neural networks |
| url | https://www.frontiersin.org/articles/10.3389/fninf.2023.1272791/full |
| work_keys_str_mv | AT thomastveitstøl introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT thomastveitstøl introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT matstveter introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT matstveter introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT anasperezt introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT anasperezt introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT christofferhatlestadhall introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT anisyazidi introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT hugolhammer introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT hugolhammer introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels AT irarjheboldharaldsen introducingregionbasedpoolingforhandlingavariednumberofeegchannelsfordeeplearningmodels |