Biomarkers.

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Alzheimer's Disease (AD) is characterized by brain dysconnectivity, but information exchange patterns using cross-time coherence mapping (CTCM) are underexplored.
• This study investigates CTCM connectivity (CTCM_conn) and gradient analysis in normal controls (NC), mild cognitive impairment (MCI), and AD groups.

Purpose of the Study:

• To investigate inter-relationships among 300 brain regions using CTCM and gradient analysis in NC, MCI, and AD.
• To identify differences in brain network connectivity across disease stages.
• To evaluate the potential of CTCM gradients as biomarkers for AD classification.

Main Methods:

• Resting-state fMRI (rsfMRI) data from 40 NC, 38 MCI, and 40 AD participants were analyzed.
• CTCM connectivity matrices were computed, and gradient analysis was performed using the BrainSpace toolbox.
• Machine learning models incorporating CTCM gradients, age, and gender were used for classification.

Main Results:

• Significant differences in CTCM gradients were found between NC and MCI (visual and Salience/Ventral Attention networks).
• Differences were also observed between MCI and AD (visual, sensorimotor, Salience/Ventral Attention, Default Mode, and dorsal attention networks), and between NC and AD (sensorimotor, dorsal attention, Limbic, and Default Mode networks).
• Machine learning models, particularly logistic regression, achieved over 90% accuracy in classifying disease stages when using CTCM gradients and demographic data.

Conclusions:

• CTCM gradients reveal progressive dysconnectivity in AD, with notable alterations in Default Mode, Dorsal Attention, Salience, and Limbic networks.
• CTCM gradients, combined with demographic features, show significant potential as biomarkers for AD progression and classification.
• This approach offers novel insights for AD diagnosis and understanding disease mechanisms.