Alzheimer's Imaging Consortium.

Area of Science:

• Neuroscience
• Medical Imaging
• Alzheimer's Disease Research

Background:

• Neuropsychiatric symptoms (NPS) are common in Alzheimer's disease (AD) and may arise from different pathological processes than cognitive decline.
• The interplay between amyloid, tau, and neurodegeneration in relation to NPS is not well understood.
• This study utilizes advanced neuroimaging techniques to investigate these relationships.

Purpose of the Study:

• To identify multi-modal neuroimaging patterns associated with affective symptoms and hyperactivity in Alzheimer's disease.
• To explore the distinct amyloid-tau-neurodegeneration (A-T-N) signatures underlying different NPS subtypes.
• To validate the biological relevance of identified neuroimaging patterns.

Main Methods:

• Utilized data from the Alzheimer's Disease Neuroimaging Initiative (ADNI-3), including amyloid PET, tau PET, and structural MRI.
• Employed a semi-supervised fusion framework (SuperBigFlica) to integrate multi-modal neuroimaging data and predict NPS.
• Validated findings using transfer learning to predict age and cognitive decline (CDR-SOB).

Main Results:

• The model successfully predicted affective symptoms (r=0.34) and hyperactivity (r=0.38) in an independent test set.
• Affective symptoms were associated with default mode network (DMN) disruptions, elevated amyloid and tau, and reduced cortical thickness.
• Hyperactivity correlated with amyloid deposition and gray matter/surface area reductions in DMN and frontoparietal networks; cingulate atrophy linked to disinhibition and agitation.

Conclusions:

• Distinct A-T-N signatures are identified for NPS subtyping in Alzheimer's disease.
• These neuroimaging patterns demonstrate clinical relevance and predictive utility for NPS.
• Future research should focus on longitudinal validation across diverse datasets.