Wavelet analysis for detecting body-movement artifacts in optical topography signals

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Optical topography (OT) is a noninvasive brain imaging technique crucial for infant studies.
• Body movements in infants frequently cause artifacts in OT signals, compromising data integrity.
• Accurate artifact detection is essential for reliable hemodynamic response measurements in pediatric neuroscience.

Purpose of the Study:

• To develop and validate a wavelet-based algorithm for automatic detection of body-movement artifacts in optical topography signals.
• To enhance the reliability of OT data by effectively eliminating movement-corrupted blocks.
• To improve the accuracy of hemodynamic response measurements in infant OT studies.

Main Methods:

• A wavelet transform was applied to segment and analyze OT signals from nine healthy infants during speech tasks.
• Signals were divided into 30-second blocks and classified as 'movement' or 'non-movement' based on video observation.
• A Monte Carlo analysis determined optimal parameters (scale=9, threshold=43) for the wavelet artifact detection algorithm.

Main Results:

• The developed wavelet method achieved a high discrimination rate of 86.3% for actual body movement.
• This performance was significantly better than a previous method, which had a discrimination rate of 80.6%.
• The wavelet-based approach demonstrated consistent and reliable artifact detection across different participants.

Conclusions:

• The wavelet-based method provides a practical and effective solution for identifying and removing body-movement artifacts in infant OT signals.
• This technique is crucial for improving the quality and reliability of data in optical topography studies involving infants.
• The findings support the broader application of this method to enhance pediatric brain imaging research.