DYNAMIC OPTIMIZED PRIORS FOR D-BAR RECONSTRUCTIONS OF HUMAN VENTILATION USING ELECTRICAL IMPEDANCE TOMOGRAPHY

Area of Science:

• Electrical Impedance Tomography
• Medical Imaging
• Computational Electromagnetics

Background:

• Accurate reconstruction of thoracic impedance is crucial for monitoring physiological processes.
• Existing 2-D D-bar reconstruction methods may lack dynamic adaptability to physiological changes.
• Time-difference imaging requires robust algorithms to handle temporal variations in conductivity.

Purpose of the Study:

• To develop and validate a novel method for incorporating dynamic spatial priors into the 2-D D-bar reconstruction algorithm.
• To enhance the accuracy of time-difference reconstructions for human thoracic data.
• To improve the characterization of dynamic physiological processes like respiration and cardiac activity.

Main Methods:

• Integration of dynamic spatial priors into the 2-D D-bar reconstruction algorithm.
• Optimization of conductivity values using scattering transform at each reconstruction frame.
• Guidance of dynamic prior updates via principal component analysis (PCA) of thoracic data.
• Application to time-difference reconstructions of human subject data.

Main Results:

• Demonstrated effectiveness of the dynamic spatial prior method on human subject ventilatory data.
• Successful updating of conductivity values based on physiological cycle timepoints.
• Improved reconstruction accuracy for dynamic thoracic impedance imaging.

Conclusions:

• The proposed method effectively incorporates dynamic spatial priors for enhanced 2-D D-bar reconstructions.
• Principal component analysis aids in capturing relevant temporal dynamics for prior updates.
• This approach shows promise for improved non-invasive monitoring of thoracic physiological processes.