Brain-computer interface semantic decoding system and method based on multi-modal cognitive state evaluation

CN122310084APending Publication Date: 2026-06-30SHANGHAI SHULI INTELLIGENT TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHANGHAI SHULI INTELLIGENT TECH CO LTD
Filing Date
2026-04-28
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing language brain-computer interface technologies face challenges in achieving deep semantic understanding and ensuring system robustness and stability. The high instability of neural signals and the lack of analysis of higher cognitive states result in insufficient decoding reliability and coarse semantic decoding granularity.

Method used

A brain-computer interface semantic decoding system based on multimodal cognitive state assessment is adopted. Through cross-modal comparison pre-training of EEG signals and text semantic features, combined with a personalized decoding model, the system uses sequence encoders and sequence decoders to achieve cross-modal spatial alignment of EEG feature sequences with user-associated semantic feature vectors, thereby improving the reliability and robustness of decoding.

Benefits of technology

It effectively suppresses the effects of physiological state fluctuations and electromagnetic interference, achieves natural and fluent semantic communication, and simultaneously quantifies the user's subjective state, reducing hardware and software deployment costs and expanding the application scenarios of language brain-computer interfaces.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122310084A_ABST
    Figure CN122310084A_ABST
Patent Text Reader

Abstract

This invention discloses a brain-computer interface semantic decoding system and method based on multimodal cognitive state assessment, belonging to the field of brain-computer interface technology. The system includes an EEG signal acquisition module, a preprocessing module, and a semantic decoding module. The EEG signal acquisition module acquires EEG signals during a user's free association task based on preset scene materials; the preprocessing module performs sequence mapping processing on the EEG signals to obtain corresponding EEG feature sequences; the semantic decoding module inputs the EEG feature sequences into a trained personalized decoding model. This model is pre-trained through cross-modal comparison of EEG signals and text semantic features, and trained in conjunction with user subjective cognitive state constraints, and includes a sequence encoder and a decoder. The model achieves cross-modal spatial alignment of the EEG feature sequences and the associative semantic feature vectors, outputting the semantic decoding result. This invention can effectively improve the accuracy of EEG semantic decoding, system robustness, and personalized adaptability.
Need to check novelty before this filing date? Find Prior Art