A PCA Based Artifact Removal Algorithm for Neural Signal Acquisition with kS/s Sampling Rate.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Deep Brain Stimulation (DBS) is a crucial therapy, but stimulus artifacts contaminate neural recordings.
• Artifact contamination poses significant challenges for accurate signal processing, particularly with non-integer sampling rate to stimulation frequency ratios.

Purpose of the Study:

• To develop and validate a novel method for eliminating stimulus artifacts in DBS neural signals.
• To enable reliable signal processing in DBS even with low neural signal sampling rates.

Main Methods:

• A transfer function was developed to model the relationship between stimulation signals and acquisition site artifacts.
• A Principal Component Analysis (PCA)-based linear regression algorithm was implemented for artifact removal.
• A numerical recipe was proposed to optimize the algorithm's computational complexity.

Main Results:

• The PCA-based algorithm effectively removed stimulus artifacts from neural signals.
• High correlation coefficients (over 60%) were achieved for artifact-free signals, even when artifacts were 60dB stronger than the neural signal.
• The numerical recipe reduced the algorithm's computational complexity from cubic to square degree.

Conclusions:

• The proposed artifact removal method is effective for DBS signal processing, especially at low sampling rates.
• The optimized algorithm offers a computationally efficient solution for real-time artifact elimination in DBS.
• This work significantly enhances the quality of neural data recorded during DBS.