Precise MRI slice extraction along the anterior commissure–posterior commissure (AC-PC) axis: a standardized approach

Abstract

Early and accurate detection of brain changes is crucial for diagnosing and understanding neurological conditions such as Alzheimer’s disease. However, differences in the orientation of brain scans can make it difficult to compare images across individuals. To address this, we present an automated method that extracts MRI slices aligned with the Anterior Commissure–Posterior Commissure (AC–PC) line, a widely accepted reference plane in neuroimaging. Our pipeline includes alignment to a standard brain template, skull stripping, and a novel slice identification technique that combines constant and dynamic thresholding to detect key anatomical landmarks. A key strength of our method is its adaptability across different Alzheimer’s diagnosis groups. This standardized extraction process improves consistency in image orientation, supporting more accurate and comparable analyses. Validation using 100 brain scans from the ADNI dataset across five diagnostic categories showed strong agreement with manual selections, with an average cross-correlation coefficient of 0.9609, an average structural similarity index measure of 0.8125, an average low mean squared error of 0.00495, and an average joint histogram coefficient of 0.8029, confirming high anatomical and intensity fidelity. By reducing preprocessing time and improving consistency, our method provides a practical and reliable solution for large-scale neuroimaging studies and the early detection of Alzheimer’s related brain conditions.