Paper List

期刊: ArXiv Preprint
发布日期: 2026-03-03
NeuroscienceBioinformatics

Toward Robust, Reproducible, and Widely Accessible Intracranial Language Brain-Computer Interfaces: A Comprehensive Review of Neural Mechanisms, Hardware, Algorithms, Evaluation, Clinical Pathways and Future Directions

The Hong Kong Polytechnic University | Chongqing University of Technology

Dongyi He, Wai Ting Siok, Nizhuan Wang
Figure
Figure
Figure
Figure
Figure

30秒速读

IN SHORT: This review addresses the core challenge of fragmented and heterogeneous evidence that hinders the clinical translation of intracranial language BCIs, providing a unified framework to bridge neuroscience, hardware, algorithms, and clinical deployment.

核心创新

  • Methodology Proposes an end-to-end, decision-oriented synthesis linking neural representations to recording choices, experimental design, decoding architectures, and translational constraints.
  • Methodology Introduces a structured framework organized around five coupled design questions and a unified evaluation framework with cross-linguistic, cross-task benchmark templates.
  • Biology Synthesizes neural mechanisms underlying overt, mimed, and imagined speech, highlighting the somatotopic organization and intermixed tuning in sensorimotor cortex, and the gradient of signal-to-noise ratio (SNR) across speech modalities.

主要结论

  • Intracranial recordings (MEA, ECoG, SEEG) enable high-performance decoding, with state-of-the-art systems achieving up to 90.9 words per minute (wpm) and Word Error Rates (WER) as low as 3% in participant-specific tasks, yet cross-subject transfer remains a major bottleneck.
  • Articulatory intermediate representations and language-prior-assisted frameworks (e.g., transformers) enhance robustness and data efficiency, with studies reporting mel-spectrogram correlation PCC ~0.806-0.838 and improved generalization in multi-subject training.
  • Clinical translation requires addressing long-term stability (e.g., median accuracy ~90.59% over 3 months without recalibration in one study) and establishing unified evaluation metrics that integrate objective, perceptual, expressive, and longitudinal measures.
研究空白： The field suffers from fragmented evidence, poor cross-subject generalization (many decoders remain subject-specific), heterogeneous evaluation practices, limited naturalistic expressivity (especially for tonal languages), and unresolved design trade-offs across hardware, algorithms, and clinical integration.

摘要: Intracranial language brain-computer interfaces (BCIs) offer a promising route for restoring communication in individuals with severe motor and speech impairments, but clinical translation remains limited by fragmented and heterogeneous evidence, as well as unresolved design trade-offs across neuroscience, hardware, algorithms, validation methods, and clinical integration. This review synthesizes recent progress across four key domains in intracranial speech neuroprosthetics: i) the neural mechanisms underlying overt, mimed, and imagined speech; ii) decision-oriented hardware comparisons of surgically implanted recording modalities, including microelectrode array (MEA), electrocorticography (ECoG), and stereotactic electroencephalography (SEEG); iii) experimental strategies for achieving cross-subject and multilingual generalization; and iv) advances in neural decoding, including sequence models, attention-based architectures (e.g., transformers), articulatory intermediate representations, and language-prior-assisted frameworks. We highlight persistent bottlenecks, including weak cross-subject transfer, long-term non-stationarity and recalibration burden, heterogeneous and non-comparable evaluation practices, limited naturalistic expressivity (especially for tonal/logosyllabic languages), and the low signal-to-noise ratio (SNR) of neural activity in covert speech decoding. Our contributions are threefold: (1) an end-to-end, decision-oriented synthesis that links neural representations to recording choices, experimental design, decoding model architectures, and translational constraints; (2) a structured framework organized around five coupled design questions, accompanied by a unified evaluation framework and a cross-linguistic, cross-task benchmark template that integrates objective, perceptual, expressive, conversational, and longitudinal metrics; and (3) user-centered translational guidance that includes agency-preserving shared control, verifiable performance priorities, and scenario-specific minimum viable system (MVP) profiles for differentiating between reliability-first home communication and fidelity-first conversational speech restoration. We conclude with a call for larger multilingual, multi-center longitudinal datasets; harmonized benchmarks; adaptive yet interpretable decoders; prospective clinical validation; and transparent data-sharing and reporting practices with robust ethical safeguards. These efforts are essential to accelerate the safe and equitable deployment of speech neuroprostheses.