Iterative N-way partial least squares for a binary self-paced brain-computer interface in freely moving animals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interface (BCI) systems require accurate calibration for reliable operation.
• Existing calibration methods may lack robustness or require extensive user input.
• Self-paced BCI systems need efficient methods to decode neural signals for control.

Purpose of the Study:

• To develop a novel tensor-based approach for calibrating binary self-paced BCI systems.
• To enable automatic extraction of BCI-related features and predictive model construction.
• To validate the proposed method in freely moving animals under naturalistic conditions.

Main Methods:

• Utilized electrocorticograms (ECoG) from freely moving rats during behavioral experiments.
• Mapped ECoG data to a spatial-temporal-frequency space using continuous wavelet transformation to form a feature tensor.
• Applied N-way partial least squares (NPLS) for tensor factorization and movement intention prediction.
• Developed an iterative NPLS (INPLS) algorithm to handle high-dimensional feature tensors.

Main Results:

• The INPLS algorithm demonstrated high accuracy and robustness in computational experiments.
• The proposed method allows for fully automatic BCI system calibration.
• The approach successfully extracts BCI-relevant features without requiring prior neurophysiological knowledge.
• A predictive model of control was constructed based on pre-BCI event neural activity.

Conclusions:

• The developed tensor-based approach offers an effective and automated solution for BCI system calibration.
• INPLS provides a robust method for analyzing high-dimensional neural data in BCI applications.
• The study validates the BCI system's performance in realistic experimental settings, paving the way for practical applications.