An iterative approach to the solution of the inverse problem

Area of Science:

• Biophysics
• Computational Neuroscience
• Medical Imaging

Background:

• The bioelectric inverse problem seeks to determine internal bioelectric sources from external measurements.
• Existing methods often struggle with physiological plausibility and measurement noise.
• A robust framework is needed to accurately reconstruct neural activity.

Purpose of the Study:

• To develop and validate an iterative approach for solving the bioelectric inverse problem.
• To incorporate physiological plausibility and measurement noise into the solution space.
• To test the efficacy of the method in a realistic head model.

Main Methods:

• Framing the bioelectric inverse problem as a search within a feasible solution set.
• Defining the feasible set to account for measurement noise.
• Employing a regularized inverse at each iterative step to constrain the search space.
• Validating the approach using a digitized cadaver cortex within a spherical head model.

Main Results:

• The iterative approach successfully narrows down the search for physiologically plausible solutions.
• Regularization effectively limits the computational search space in each iteration.
• The model system demonstrated the practical application of the developed method.

Conclusions:

• The proposed iterative method offers a robust solution to the bioelectric inverse problem.
• Accounting for noise and ensuring physiological plausibility are critical for accurate source localization.
• This technique holds promise for improved neuroimaging and brain activity reconstruction.