Automatic identification of gray and white matter components in polarized light imaging

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Polarized light imaging (PLI) visualizes brain fiber tracts at high resolution.
• PLI data analysis derives fiber orientation vectors from optical signals.
• Noise and artifacts in PLI signals impact fiber tracking accuracy.

Purpose of the Study:

• To develop a user-independent method for identifying gray and white matter components in PLI data.
• To enable automated, large-scale analysis of nerve fiber orientations in the human brain.

Main Methods:

• Utilized independent component analysis (ICA) for artifact rejection and signal restoration in PLI data.
• Developed an automatic method based on statistical properties of component feature vectors to identify gray and white matter components.
• Applied the method to analyze nerve fiber orientations in thousands of whole human brain sections.

Main Results:

• Successfully restored and enhanced sinusoidal PLI signals, particularly in low-intensity regions.
• Achieved user-independent identification of gray and white matter components from decomposed PLI data.
• Demonstrated the capability for large-scale analysis of human brain nerve fiber orientations.

Conclusions:

• The proposed automatic method effectively extracts relevant signals from complex PLI data.
• This approach facilitates robust and scalable analysis of brain connectivity using PLI.
• Enables comprehensive mapping of nerve fiber orientations across extensive human brain datasets.