Extracting task-related activation components from optical topography measurement using independent components...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Optical topography (OT) measures brain activity by detecting blood oxygenation changes.
• OT signals can be confounded by physiological changes unrelated to neuronal activation.
• Existing methods often rely on hemodynamic models, which may not fully capture signal dynamics.

Purpose of the Study:

• To develop and validate a method for separating neuronal-activation signals from physiological noise in OT data.
• To extract brain activation components without relying on predefined hemodynamic models.
• To apply the developed method to analyze OT signals from motor tasks.

Main Methods:

• Independent Component Analysis (ICA) was employed to decompose OT signals.
• A time-delayed decorrelation algorithm separated signals into independent components (ICs).
• Task-related ICs (TR-ICs) were identified by cross-correlations, then clustered into activation-related (TR-AICs) and noise (TR-NICs) using k-means.

Main Results:

• The ICA-based method successfully separated brain activation signals from noise in OT data.
• Task-related activation components (TR-AICs) exhibited distinct waveform patterns, including overshoots post-task.
• Noise components (TR-NICs) showed a characteristic N-shaped waveform, differentiating them from true activation.

Conclusions:

• Independent component analysis provides a robust approach to isolate neuronal activation in optical topography.
• This model-free technique enhances the accuracy of brain function analysis from OT measurements.
• The identified waveform patterns offer potential biomarkers for distinguishing neural activity from physiological artifacts.