MEGConformer: Conformer-Based MEG Decoder for Robust Speech and Phoneme Classification
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918329890897920 |
|---|---|
| author | de Zuazo, Xabier Saratxaga, Ibon Navas, Eva |
| author_facet | de Zuazo, Xabier Saratxaga, Ibon Navas, Eva |
| contents | Decoding speech-related information from non-invasive MEG is a key step toward scalable brain-computer interfaces. We present compact Conformer-based decoders on the LibriBrain 2025 PNPL benchmark for two core tasks: Speech Detection and Phoneme Classification. Our approach adapts a compact Conformer to raw 306-channel MEG signals, with a lightweight convolutional projection layer and task-specific heads. For Speech Detection, a MEG-oriented SpecAugment provided a first exploration of MEG-specific augmentation. For Phoneme Classification, we used inverse-square-root class weighting and a dynamic grouping loader to handle 100-sample averaged examples. In addition, a simple instance-level normalization proved critical to mitigate distribution shifts on the holdout split. Using the official Standard track splits and F1-macro for model selection, our best systems achieved 88.9% (Speech) and 65.8% (Phoneme) on the leaderboard, winning the Phoneme Classification Standard track. For further implementation details, the technical documentation, source code, and checkpoints are available at https://github.com/neural2speech/libribrain-experiments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_01443 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MEGConformer: Conformer-Based MEG Decoder for Robust Speech and Phoneme Classification de Zuazo, Xabier Saratxaga, Ibon Navas, Eva Computation and Language Machine Learning Neural and Evolutionary Computing Sound 68T07 (Primary), 92C55, 62H30 (Secondary) I.2.6; I.5.1; J.3 Decoding speech-related information from non-invasive MEG is a key step toward scalable brain-computer interfaces. We present compact Conformer-based decoders on the LibriBrain 2025 PNPL benchmark for two core tasks: Speech Detection and Phoneme Classification. Our approach adapts a compact Conformer to raw 306-channel MEG signals, with a lightweight convolutional projection layer and task-specific heads. For Speech Detection, a MEG-oriented SpecAugment provided a first exploration of MEG-specific augmentation. For Phoneme Classification, we used inverse-square-root class weighting and a dynamic grouping loader to handle 100-sample averaged examples. In addition, a simple instance-level normalization proved critical to mitigate distribution shifts on the holdout split. Using the official Standard track splits and F1-macro for model selection, our best systems achieved 88.9% (Speech) and 65.8% (Phoneme) on the leaderboard, winning the Phoneme Classification Standard track. For further implementation details, the technical documentation, source code, and checkpoints are available at https://github.com/neural2speech/libribrain-experiments. |
| title | MEGConformer: Conformer-Based MEG Decoder for Robust Speech and Phoneme Classification |
| topic | Computation and Language Machine Learning Neural and Evolutionary Computing Sound 68T07 (Primary), 92C55, 62H30 (Secondary) I.2.6; I.5.1; J.3 |
| url | https://arxiv.org/abs/2512.01443 |