Cross entropy gradient analysis for Alzheimer's disease assessment

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Alzheimer's disease (AD) is characterized by cognitive decline due to disrupted brain region coordination.
• Cross-entropy quantifies inter-regional information flow, but its application in AD is unexplored.

Purpose of the Study:

• To evaluate cross-regional brain entropy (CRBEN) in normal controls (NC), mild cognitive impairment (MCI), and AD using resting-state fMRI.
• To assess CRBEN's potential as a biomarker for Alzheimer's disease progression.

Main Methods:

• Resting-state fMRI data from 40 NC, 38 MCI, and 40 AD participants were analyzed.
• CRBEN matrices were computed and decomposed into functional gradients.
• Machine learning classifiers were trained to assess diagnostic accuracy.

Main Results:

• MCI and AD patients exhibited reduced CRBEN gradients compared to NC.
• AD showed further gradient reductions in specific brain networks compared to MCI and NC.
• Machine learning models achieved approximately 90% accuracy in classifying disease stages.

Conclusions:

• CRBEN gradients indicate progressive loss of network complexity across the Alzheimer's disease continuum.
• CRBEN may serve as a sensitive biomarker for detecting functional disintegration in AD.
• Functional gradients derived from CRBEN offer insights into the neurobiological underpinnings of AD progression.