Characterization of Fornix Structural Changes in Controls, MCI, and AD Using sMR Images and Rotational Invariant Moments

Alzheimer's disease (AD) is a prevalent neurodegenerative form of dementia that progressively affects individuals. Alterations in the fornix region are significant biomarkers of AD. In this letter, an approach is proposed to distinguish mild cognitive impairment (MCI) from normal cognition (NC) and AD subjects using structural magnetic resonance (sMR) images and rotational invariant moments (RIMs). Initially, the sMR images are preprocessed using a standard pipeline, and the fornix structure is segmented using the reaction-diffusion level set (RDLS) algorithm. Then, a total of 11 RIMs are computed from the fornix region. Further, the significant features are identified using statistical tests: Shapiro–Wilk, Wilcoxon's rank-sum, and one-way analysis of variance. The results indicate that RDLS effectively segments the fornix structure. The RIM features demonstrate high statistical significance (<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> < 0.01) in differentiating MCI from NC and AD, highlighting their ability to capture shape variations in the fornix. The proposed approach effectively differentiates MCI from NC and AD subjects, suggesting its clinical relevance in Alzheimer's condition analysis.