Iterative refinement of the minimum norm solution of the bioelectric inverse problem

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Functional brain imaging relies on solving the inverse problem to estimate cortical current from scalp potentials.
• Current minimum norm least squares solutions provide a unique but often widespread distribution of cortical activity.
• Physiologically relevant solutions may involve restricted regions of cortical current generation.

Purpose of the Study:

• To develop a novel method for uncovering physiologically relevant, restricted cortical current distributions.
• To improve the localization accuracy of cortical activity in functional brain imaging.
• To account for the influence of measurement noise on feasible solutions.

Main Methods:

• Iterative application of the minimum norm approach to define a shrinking region of interest (ellipsoid).
• Searching for feasible solutions within the iteratively refined region.
• Incorporating measurement noise explicitly in the solution process.
• Testing the method using a realistic three-shell sphere head model with a detailed cortex.

Main Results:

• The proposed method successfully uncovers feasible minimum norm solutions with restricted cortical current generation.
• The iterative ellipsoid shrinkage effectively guides the search towards the active cortical region.
• The method demonstrates improved localization of simulated cortical activity compared to standard approaches.
• Explicit consideration of noise enhances the reliability of identifying localized activity.

Conclusions:

• The novel method enhances the precision of functional brain imaging by identifying localized cortical activity.
• This approach offers a more physiologically plausible interpretation of brain activity compared to widespread solutions.
• The technique holds promise for advancing our understanding of brain function and neurological disorders.