mlsanityData to 13 Oct 2026

SHINE: Sequential Hierarchical Integration Network for EEG and MEG

arXiv 2602.23960 ↗First posted 27 Feb 2026cs.SD cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

How natural speech is represented in the brain constitutes a major challenge for cognitive neuroscience. Reconstructing the speech envelope and Mel spectrogram from EEG and MEG provides a time-resolved way to study its temporal and spectral structure. Speech-related neural activity spans sensors and temporal scales; extracting these representations while adapting the use of context to each acoustic target is a central problem in speech reconstruction. We propose SHINE, a Sequential Hierarchical Integration Network for EEG and MEG. A residual sensor adapter unifies input dimensions, intermediate dilated-block states retain temporal depth, and a target- and time-dependent gate fuses local hierarchical and attention-enhanced context predictions. Across two EEG and two MEG datasets, SHINE has the highest mean envelope and mean-Mel Pearson correlations among nine local baseline implementations on all eight dataset-metric combinations. SHINE also placed second in the speech-detection Extended Track of the NeurIPS 2025 PNPL Competition. Code will be released at https://github.com/xuxiran/SHINE.

Key facts. Cited by 1 paper in the map (data through 13 Oct 2026). Builds on 16 of its 43 references that are in the map. First posted on arXiv 27 Feb 2026.

Builds on

16 of 43 references.

References · 16

  1. 2022Decoding speech perception from non-invasive brain recordingscited by 129
  2. 2025LibriBrain: Over 50 Hours of Within-Subject MEG to Improve Speech Decoding Methods at Scalecited by 20
  3. 2025The 2025 PNPL Competition: Speech Detection and Phoneme Classification in the LibriBrain Datasetcited by 16
  4. 2024ConvConcatNet: a deep convolutional neural network to reconstruct mel spectrogram from the EEGcited by 12
  5. 2023Relating EEG to continuous speech using deep neural networks: a reviewcited by 11
  6. 2024DARNet: Dual Attention Refinement Network with Spatiotemporal Construction for Auditory Attention Detectioncited by 10
  7. 2023A DenseNet-based method for decoding auditory spatial attention with EEGcited by 8
  8. 2021Convolutional Neural Networks Demystified: A Matched Filtering Perspective Based Tutorialcited by 6
  9. 2024Beware of Overestimated Decoding Performance Arising from Temporal Autocorrelations in Electroencephalogram Signalscited by 6
  10. 2025SSM2Mel: State Space Model to Reconstruct Mel Spectrogram from the EEGcited by 5
  11. 2021Auditory Attention Decoding from EEG using Convolutional Recurrent Neural Networkcited by 4
  12. 2024Self-supervised speech representation and contextual text embedding for match-mismatch classification with EEG recordingcited by 3
4 more
  1. 2025DMF2Mel: A Dynamic Multiscale Fusion Network for EEG-Driven Mel Spectrogram Reconstructioncited by 3
  2. 2023HappyQuokka System for ICASSP 2023 Auditory EEG Challengecited by 2
  3. 2026DECAF: Dynamic Envelope Context-Aware Fusion for Speech-Envelope Reconstruction from EEGcited by 2
  4. 2024StreamAAD: Decoding Spatial Auditory Attention with a Streaming Architecturecited by 1

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026MindVoice: Reconstructing Intelligible Speech from Non-invasive Neural Signals with Pretrained Priorscited by 0