A Review of EEG-Based Speech Imagery Encoding Paradigms and Decoding Algorithms
Journal of Brain-Computer Interface · 01.10.2025 · 1 (1) · 58
DOI: 10.58195/bci.2025.1.008 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Brain-Computer Interface, online erschienen am 01.10.2025. Verfasst von Bang-hua Yang und 3 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: 国家自然科学基金资助项目.
Abstract (Originalsprache)
Electroencephalogram (EEG) provides the core decoding foundation for speech imagery brain-computer interfaces (BCI), enabling individuals with speech impairments to achieve human-computer interaction without relying on peripheral nerves and muscle pathways. It holds significant potential in areas such as aphasia rehabilitation and covert communication. This article systematically reviews advances in neural encoding paradigms and decoding algorithms for EEG-based speech imagery BCI: In terms of encoding, the neural patterns during speech imagery highly overlap with those during speech execution, primarily involving key nodes of the language network such as Broca's area and Wernicke's area. The imagined content spans phonemes, syllables, words, and sentences. Paradigm design often employs multimodal stimuli and repetitive tasks to enhance signal separability. Regarding decoding, traditional two-stage methods (feature extraction + classification) are limited by their dependence on feature engineering and poor cross-subject generalization. End-to-end deep learning models directly process raw spatiotemporal EEG sequences, achieving joint optimization of feature extraction and classification, which significantly improves decoding efficiency. Emerging approaches integrate natural language processing and pre-trained language models, leveraging semantic context to achieve sentence-level decoding and offering new pathways for complex language reconstruction. Current key challenges include insufficient decoding accuracy and poor cross-subject generalization. Future efforts should incorporate few-shot learning frameworks to enhance model generalization, integrate multimodal data to deepen the understanding of neural mechanisms underlying EEG, and promote synergy between EEG signals and large language models to advance practical clinical applications.
Förderung
- 国家自然科学基金资助项目
Zitieren
APA
Yang, B., Gan, A., Guo, W., & Kuang, S. (2025). A Review of EEG-Based Speech Imagery Encoding Paradigms and Decoding Algorithms. Journal of Brain-Computer Interface, 1(1), 58. https://doi.org/10.58195/bci.2025.1.008
Vancouver
Yang B, Gan A, Guo W, Kuang S. A Review of EEG-Based Speech Imagery Encoding Paradigms and Decoding Algorithms. Journal of Brain-Computer Interface. 2025;1(1):58. doi:10.58195/bci.2025.1.008
BibTeX
@article{Yang2025,
author = {Yang, Bang-hua and Gan, An-an and Guo, Wen-bin and Kuang, Shuai},
title = {A Review of EEG-Based Speech Imagery Encoding Paradigms and Decoding Algorithms},
journal = {Journal of Brain-Computer Interface},
year = {2025},
volume = {1},
number = {1},
pages = {58},
doi = {10.58195/bci.2025.1.008}
}
EUVIMED