Multi-Channel Vision Transformer for Epileptic Seizure Prediction
Epilepsy is a neurological disorder that causes recurrent seizures and sometimes loss of awareness. Around 30% of epileptic patients continue to have seizures despite taking anti-seizure medication. The ability to predict the future occurrence of seizures would enable the patients to take precaution...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Biomedicines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-9059/10/7/1551 |
| _version_ | 1797433826338668544 |
|---|---|
| author | Ramy Hussein Soojin Lee Rabab Ward |
| author_facet | Ramy Hussein Soojin Lee Rabab Ward |
| author_sort | Ramy Hussein |
| collection | DOAJ |
| description | Epilepsy is a neurological disorder that causes recurrent seizures and sometimes loss of awareness. Around 30% of epileptic patients continue to have seizures despite taking anti-seizure medication. The ability to predict the future occurrence of seizures would enable the patients to take precautions against probable injuries and administer timely treatment to abort or control impending seizures. In this study, we introduce a Transformer-based approach called Multi-channel Vision Transformer (MViT) for automated and simultaneous learning of the spatio-temporal-spectral features in multi-channel EEG data. Continuous wavelet transform, a simple yet efficient pre-processing approach, is first used for turning the time-series EEG signals into image-like time-frequency representations named Scalograms. Each scalogram is split into a sequence of fixed-size non-overlapping patches, which are then fed as inputs to the MViT for EEG classification. Extensive experiments on three benchmark EEG datasets demonstrate the superiority of the proposed MViT algorithm over the state-of-the-art seizure prediction methods, achieving an average prediction sensitivity of 99.80% for surface EEG and 90.28–91.15% for invasive EEG data. |
| first_indexed | 2024-03-09T10:22:19Z |
| format | Article |
| id | doaj.art-777d2c86e311469f8afadd1c5854fe03 |
| institution | Directory Open Access Journal |
| issn | 2227-9059 |
| language | English |
| last_indexed | 2024-03-09T10:22:19Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Biomedicines |
| spelling | doaj.art-777d2c86e311469f8afadd1c5854fe032023-12-01T21:55:51ZengMDPI AGBiomedicines2227-90592022-06-01107155110.3390/biomedicines10071551Multi-Channel Vision Transformer for Epileptic Seizure PredictionRamy Hussein0Soojin Lee1Rabab Ward2Center for Advanced Functional Neuroimaging, Stanford University, Stanford, CA 94305, USAPacific Parkinson’s Research Centre, University of British Columbia, Vancouver, BC V6T 2B5, CanadaElectrical and Computer Engineering, University of British Columbia, Vancouver, BC V6T 1Z4, CanadaEpilepsy is a neurological disorder that causes recurrent seizures and sometimes loss of awareness. Around 30% of epileptic patients continue to have seizures despite taking anti-seizure medication. The ability to predict the future occurrence of seizures would enable the patients to take precautions against probable injuries and administer timely treatment to abort or control impending seizures. In this study, we introduce a Transformer-based approach called Multi-channel Vision Transformer (MViT) for automated and simultaneous learning of the spatio-temporal-spectral features in multi-channel EEG data. Continuous wavelet transform, a simple yet efficient pre-processing approach, is first used for turning the time-series EEG signals into image-like time-frequency representations named Scalograms. Each scalogram is split into a sequence of fixed-size non-overlapping patches, which are then fed as inputs to the MViT for EEG classification. Extensive experiments on three benchmark EEG datasets demonstrate the superiority of the proposed MViT algorithm over the state-of-the-art seizure prediction methods, achieving an average prediction sensitivity of 99.80% for surface EEG and 90.28–91.15% for invasive EEG data.https://www.mdpi.com/2227-9059/10/7/1551EEGepilepsyseizure predictioncontinuous wavelet transformvision transformer |
| spellingShingle | Ramy Hussein Soojin Lee Rabab Ward Multi-Channel Vision Transformer for Epileptic Seizure Prediction Biomedicines EEG epilepsy seizure prediction continuous wavelet transform vision transformer |
| title | Multi-Channel Vision Transformer for Epileptic Seizure Prediction |
| title_full | Multi-Channel Vision Transformer for Epileptic Seizure Prediction |
| title_fullStr | Multi-Channel Vision Transformer for Epileptic Seizure Prediction |
| title_full_unstemmed | Multi-Channel Vision Transformer for Epileptic Seizure Prediction |
| title_short | Multi-Channel Vision Transformer for Epileptic Seizure Prediction |
| title_sort | multi channel vision transformer for epileptic seizure prediction |
| topic | EEG epilepsy seizure prediction continuous wavelet transform vision transformer |
| url | https://www.mdpi.com/2227-9059/10/7/1551 |
| work_keys_str_mv | AT ramyhussein multichannelvisiontransformerforepilepticseizureprediction AT soojinlee multichannelvisiontransformerforepilepticseizureprediction AT rababward multichannelvisiontransformerforepilepticseizureprediction |