| DOI | Resolve DOI: https://doi.org/10.1109/SENSORS60989.2024.10784646 |
|---|
| Author | Search for: Cook, Zara1ORCID identifier: https://orcid.org/0009-0009-6059-4377; Search for: Zhao, Chengzong1; Search for: Murray, Livia1; Search for: Kesan, Jivan1; Search for: Belacel, Nabil1ORCID identifier: https://orcid.org/0000-0003-1731-3225; Search for: Doesburg, Sam M.; Search for: Medvedev, George; Search for: Vakorin, Vasily A.; Search for: Xi, Pengcheng1ORCID identifier: https://orcid.org/0000-0003-3236-5234 |
|---|
| Affiliation | - National Research Council Canada. Digital Technologies
|
|---|
| Format | Text, Article |
|---|
| Conference | 2024 IEEE SENSORS, October 20-23, 2024, Kobe, Japan |
|---|
| Subject | EEG analysis; brain age prediction; self-supervised learning; Graph Neural Networks |
|---|
| Abstract | Electroencephalogram (EEG) recordings are valuable for capturing neuro-physiological states, with brain age prediction providing key insights into brain health. To scale this diagnostic technique, we propose a computer-aided system using self-supervised learning (SSL) and Graph Neural Networks (GNNs) for EEG analysis. SSL reduces the need for fully labeled data by pre-training models on large unlabeled EEG datasets. We tackle temporal-spectral feature learning challenges with GNNs, employing graph-based representations of EEG data to depict the brain's interconnectedness and extract meaningful features. Furthermore, we enhance the explainability of brain age predictions by visualizing channel-wise maps, highlighting critical EEG channels Influencing the model's decisions. |
|---|
| Publication date | 2024-12-17 |
|---|
| Publisher | IEEE |
|---|
| In | |
|---|
| Language | English |
|---|
| Peer reviewed | Yes |
|---|
| Export citation | Export as RIS |
|---|
| Report a correction | Report a correction (opens in a new tab) |
|---|
| Record identifier | adb997d9-e5ee-4bea-a769-d30854a33e37 |
|---|
| Record created | 2024-12-20 |
|---|
| Record modified | 2024-12-23 |
|---|
