Improved Motion Artifact Correction in fNIRS Data by Combining Wavelet and Correlation-Based Signal Improvement

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique enabling participant movement.
• Head movements during fNIRS acquisition introduce motion artifacts (MA) that contaminate the neural signal.
• Effective MA correction is crucial for reliable fNIRS data interpretation.

Purpose of the Study:

• To introduce and evaluate an improved algorithmic approach for motion artifact correction in fNIRS.
• To compare the proposed wavelet and correlation-based signal improvement (WCBSI) method against established MA correction techniques.
• To assess the MA correction accuracy using quantitative metrics on real fNIRS data.

Main Methods:

• Developed a novel algorithm combining wavelet and correlation-based signal improvement (WCBSI).
• Collected fNIRS data from 20 participants performing a hand-tapping task with induced head movements.
• Compared WCBSI against seven established methods (spline interpolation, Savitzky-Golay, PCA, tPCA, Loess, wavelet, correlation-based) using R, RMSE, MAPE, and ΔAUC metrics.

Main Results:

• The WCBSI algorithm demonstrated superior performance in correcting motion artifacts across all tested metrics.
• WCBSI was the only algorithm to significantly exceed average performance (p < 0.001).
• The proposed WCBSI approach showed a 78.8% probability of being the top-ranked algorithm.

Conclusions:

• The WCBSI algorithm offers a robust and effective solution for motion artifact correction in fNIRS.
• This improved method enhances the reliability and accuracy of fNIRS neuroimaging data.
• WCBSI represents a significant advancement for motion artifact removal in mobile neuroimaging applications.