NeuXus open-source tool for real-time artifact reduction in simultaneous EEG-fMRI

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) offer complementary insights into brain activity.
• Real-time artifact reduction is crucial for applications like neurofeedback, but existing methods are often slow, hardware-specific, or proprietary.
• Gradient and pulse artifacts significantly contaminate EEG signals during fMRI acquisition.

Purpose of the Study:

• To develop and validate a fast, open-source, and hardware-independent tool for real-time EEG artifact reduction in simultaneous EEG-fMRI.
• To enable advanced neurofeedback applications by improving the quality of real-time EEG data.
• To provide a publicly available solution that overcomes limitations of current artifact reduction techniques.

Main Methods:

• Integration of artifact average subtraction with a long short-term memory network for R peak detection within the NeuXus Python toolbox.
• Real-time processing of EEG signals acquired during simultaneous EEG-fMRI recordings.
• Benchmarking against established offline and real-time artifact reduction tools using multiple datasets.

Main Results:

• NeuXus demonstrated comparable or superior artifact power reduction and background signal preservation compared to existing tools.
• The tool achieved execution times under 250 ms, confirming its real-time capability.
• Validation across resting state, alpha-band reactivity, and motor imagery tasks confirmed performance.

Conclusions:

• NeuXus is the first fully open-source and hardware-independent solution for real-time EEG artifact reduction in simultaneous EEG-fMRI.
• This tool facilitates advanced neurofeedback and other real-time applications by providing high-quality, artifact-free EEG data.
• The open-source nature of NeuXus promotes accessibility and further development in the field of multimodal neuroimaging.