Sunday, December 3, 2023

WiMi Developed a Motor Imagery Brain-computer Interface Based on Multi-source Signal Processing

Related stories


WiMi Hologram Cloud Inc., a leading global Hologram Augmented Reality (“AR”) Technology provider, announced that a motor imagery brain-computer interface (MI-BCI) based on multi-source signal processing has been developed.

WiMi’s MI-BCI development aims to overcome the challenges of traditional BCI systems, which include signal noise, poor classification accuracy, and other issues. By introducing a multi-source signal processing approach, this innovative technology enables more accurate brain signal parsing and processing, providing users with higher control accuracy and wider application potential. This technology is expected to lead to the next important milestone in the field of BCI. Its main features and key technology points:

Multi-source signal processing: The technology employs an advanced multi-source signal processing method that utilizes multiple sources of EEG signals, not just channel signals. This means it can capture and interpret brain activity more accurately, which improves the performance of the system.

Common spatial patterns (CSP): In the early stages of signal processing, CSP algorithms are applied to each sub-band to optimize the extraction of signal features. CSP is widely used in the field of BCI and helps to maximize the differentiation of different types of brain signals.

Blind source separation (BSS): The BSS is used to identify and separate unknown and independent sources in a mixed signal. This step helps to eliminate noise and artifacts and improves the reliability of the system.

ICA-based channel identification: This technology uses an algorithm based on independent component analysis (ICA) to identify and eliminate defective signal channels to reduce the impact of inefficient input signals on system performance.

Bayesian discriminant and linear discriminator based analysis (LDA) clustering algorithms: These advanced classification algorithms are used to improve the classification performance of the system, especially when dealing with human error in subjects. They help to improve the system’s ability to recognize and classify different brain signals.

WiMi brings unprecedented accuracy and stability to BCI systems. This technology will provide users with a wider range of control and interaction capabilities, which is potentially important not only for the medical field, but also opens up new possibilities in areas such as virtual reality, gaming, and smart homes. For example, people with disabilities could more easily control electronic devices, gamers could realize a more intuitive gaming experience, and researchers could study brain activity in greater depth. This technology will advance the field of brain-computer interfaces and bring great potential to a variety of application areas.

The implementation approach and system framework of WiMi’s MI-BCI based on multi-source signal processing requires in-depth technical knowledge and engineering design. Technology realization approach:

Signal acquisition: The first task is to acquire EEG signals. This can be accomplished with an electroencephalogram (EEG) electrode array, usually placed on the scalp. However, a multi-source signal processing approach will consider multiple signal sources, including EEG, functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG), etc., to capture brain activity information more comprehensively.

Also Read: Skan launches ProcessGPT™, the first suite of generative AI capabilities for process mining and execution, combining the best of human and machine intelligence

Signal preprocessing: Acquired signals often contain noise and interference and require preprocessing to clean up the data. This includes steps such as filtering, noise removal, and time/frequency domain transformations to ensure the quality of the input data.

Multi-source signal integration: Integrating signals from different sources into a unified data representation. This can be achieved by aligning and normalizing data from different signal sources for subsequent processing.

CSP: A CSP algorithm is applied to further enhance the characterization of brain signals. CSP is a supervised learning algorithm designed to maximize the distinction between brain signals with different motor imagery, thereby improving classification accuracy. CSP can be applied to every signal source.

BSS: The BSS technique is used to identify and separate unknown and independent sources in a mixed signal. This step helps to eliminate noise and artifacts, further improving the quality of the signal.

Feature extraction and selection: Features related to the motion imagery are extracted from the multi-source signals. This may include frequency domain features, time domain features, etc. Feature selection algorithms can also be used to reduce computational complexity and improve classification performance.

Classifier training and testing: A classifier is trained using a training dataset, e.g., support vector machines (SVMs), deep learning models, etc. The trained classifier can be used to map brain signals to specific motor imagery or actions.

Real-time feedback or applications: The final system could provide real-time feedback, connecting the user’s brain signals to external devices or applications. This could include controlling a smart wheelchair, movement in a virtual reality environment, game control, etc.

SOURCE: PRNewswire


- Never miss a story with notifications

    Latest stories