Asymmetric EEG Coding for Brain-Computer Interface Signal Clarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional brain-computer interfaces (BCIs) face limitations due to large stimulation areas, high intensity, and cognitive resource occupation, which hinder long-term multi-task operation and signal extraction from noisy background EEG, particularly in visual BCI systems.
Innovation Solution
An asymmetric EEG-based coding and decoding method that utilizes spatial division multiple access (SDMA), code division multiple access (CDMA), frequency division multiple access (FDMA), and phase division multiple access (PDMA) coding, combined with discriminant mode spatial filtering and template matching, to enhance signal-to-noise ratio and classification efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional visual BCI systems use strong visual stimuli with large stimulation areas, then signal strength and distinct EEG features are improved, but cognitive resource occupation and user fatigue increase
Solution Approach 1:
The patent divides the visual stimulus into multiple small elements arranged in a grid pattern, where each element independently elicits asymmetric VEP. This segmentation allows the system to achieve sufficient signal strength through multiple weak signals while reducing the overall cognitive load compared to a single large stimulus.
Solution Approach 2:
The patent utilizes asymmetric visual evoked potential (aVEP) characteristics where stimuli presented to one visual field elicit stronger responses in the contralateral hemisphere. By presenting stimuli asymmetrically and using this neural asymmetry, the system achieves better signal discrimination with smaller stimulus areas, reducing cognitive resource occupation.
2Measurement precision
If traditional filtering methods are used to eliminate interference signals, then signal clarity is improved, but frequency information and signal characteristics are lost
Solution Approach 1:
The patent extracts and utilizes the asymmetric component of the VEP signal, separating the useful asymmetric information from the symmetric background noise and interference. By focusing on the asymmetric characteristic rather than applying broad filtering, the method preserves frequency information while enhancing signal clarity.
Solution Approach 2:
The patent introduces asymmetric VEP as an intermediary feature that bridges the raw EEG signal and the final classification. This intermediary characteristic allows for signal enhancement without the need for aggressive filtering that would lose frequency information, as the asymmetry itself serves as the discrimination basis.
3Adaptability or versatility
If the number of stimulus elements is increased to expand instruction sets, then system capability is improved, but signal complexity and processing difficulty increase
Solution Approach 1:
The patent segments the stimulus into multiple independent elements, each capable of generating a distinguishable asymmetric VEP pattern. This segmentation enables the system to encode multiple instructions simultaneously through different stimulus configurations while maintaining relatively simple processing, as each element's response can be independently analyzed.
Solution Approach 2:
The patent leverages the spatial distribution of stimulus elements across the visual field as an additional dimension for encoding information. By presenting elements at different positions and utilizing the contralateral superiority effect, the system can distinguish between multiple instructions based on spatial patterns rather than increasing temporal complexity, thus expanding instruction sets without proportionally increasing processing difficulty.
Data Source
AI summary
The present invention provides an asymmetric EEG-based coding and decoding methods for BCIs, the BCI system includes an evoked stimulus module, an acquisition module and an EEG signal data set including a training set Xk and a testing sample Y, and an EEG signal decoding module; the evoked stimulus module sends a hybrid coding visual stimulus to subjects to evoke a specific EEG signal as required; the acquisition module obtains data information by amplifying and filtering the EEG signal so as to constitute EEG signal module; the decoding module coverts the data information into an instruction set for outputting the coding method of the present invention uses asymmetric characteristics of brain electrophysiological activity response to stimulus, combines with coding strategies such as SDMA, CDMA, FDMA and phase division multiple access coding.