Related Experiment Video
Updated: Jan 9, 2026
10:05
DTI of the Visual Pathway - White Matter Tracts and Cerebral Lesions
Published on: August 26, 2014
14.5K
Machine learning for white matter fibre tract visualization in the human brain via Mueller matrix polarimetric data
Richard McKinley1, Leonard A Felger2, Ekkehard Hewer3
1Support Center for Advanced Neuroimaging (SCAN), University Institute of Diagnostic and Interventional Neuroradiology, University of Bern, Inselspital, Bern University Hospital, Bern, Switzerland.
Proceedings of Spie--The International Society for Optical Engineering
|December 11, 2025
Summary
Mueller polarimetric imaging distinguishes brain tumors from healthy white matter by detecting differences in refractive index anisotropy. This technique, combined with machine learning, enables automated brain fiber tracking for safer, more complete neurosurgery.
Area of Science:
- Neuroscience
- Biomedical Optics
- Medical Imaging
Background:
- Accurate differentiation between brain tumors and healthy tissue is crucial for maximal tumor resection and neurological function preservation during surgery.
- Existing methods for tissue differentiation often face limitations in cost, speed, and precision.
- Brain white matter exhibits optical anisotropy due to aligned myelinated axons, unlike the chaotic growth of tumor cells.
Purpose of the Study:
- To investigate the potential of Mueller polarimetric imaging for distinguishing brain tumors from healthy white matter.
- To develop and validate automated brain fiber tracking algorithms using polarimetric data for neurosurgical applications.
Main Methods:
- A wide-field visible wavelength imaging Mueller polarimetric system was employed to analyze formalin-fixed human brain sections in reflection mode.
- Non-linear decomposition of Mueller matrices yielded maps of depolarization, scalar retardance, and optical axis azimuth.
- Classical computer vision and machine learning algorithms, including a convolutional neural network, were utilized for automated white matter identification and fiber tracking.
Main Results:
- A strong correlation was established between the optical axis azimuth derived from polarimetric data and the orientation of brain fiber tracts, confirmed by histology.
- The developed machine learning approach successfully identified white matter and visualized surface fiber tracts.
- Mueller polarimetric imaging demonstrated its capability to differentiate tissue types based on optical properties.
Conclusions:
- Mueller polarimetric imaging offers a promising non-invasive method for real-time tissue characterization during neurosurgery.
- The integration of machine learning algorithms enables automated visualization of white matter tracts, aiding surgical navigation.
- This approach has the potential to enhance surgical precision, improve patient outcomes by sparing critical structures, and increase the completeness of tumor resection.
