Elucidating the neuropathological and molecular heterogeneity of amyloid beta and tau in Alzheimer's disease through machine learning and transcriptomic integration.
Authors
Affiliations (12)
Affiliations (12)
- Department of Bioengineering, Lehigh University, Bethlehem, Pennsylvania, USA.
- Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, USA.
- Center for Neuroscience Research, Children's National Hospital, Washington, District of Columbia, USA.
- George Washington University School of Medicine, Washington, District of Columbia, USA.
- Penn State College of Medicine, Hershey, Pennsylvania, USA.
- Department of Psychiatry, New York University Grossman School of Medicine, New York, New York, USA.
- Sant Pau Memory Unit, Hospital de la Santa Creu i Sant Pau - Biomedical Research Institute Sant Pau, Barcelona, Spain.
- Centro de Investigación Biomédica en Red de Enfermedades Neurodegenerativas, CIBERNED, Madrid, Spain.
- Department of Psychology, Lehigh University, Bethlehem, Pennsylvania, USA.
- Center for Psychedelic Research and Therapy, Department of Psychiatry and Behavioral Sciences, Dell Medical School, The University of Texas at Austin, Austin, Texas, USA.
- Wu Tsai Neurosciences Institute, Stanford University, Stanford, California, USA.
- Stanford Institute for Human-Centered Artificial Intelligence, Stanford, California, USA.
Abstract
Functional brain network alterations associated with Alzheimer's disease (AD) pathology, including amyloid beta (Aβ) and phosphorylated tau (p-tau), are difficult to interpret due to overlapping aging-associated and non-amyloid biological processes. We analyzed resting-state functional magnetic resonance imaging (fMRI) from 289 older adults classified as Aβ-positive (A<sup>+</sup>, n = 129) or Aβ-negative (A<sup>-</sup>, n = 160) based on cerebrospinal fluid biomarkers. A contrastive deep learning framework was used to identify A<sup>+</sup>-specific network dimensions and predict individual Aβ and p-tau levels. A<sup>+</sup>-specific signatures were localized to the right superior temporal and anterior cingulate cortices and linked to attention and memory functions, with transcriptomic enrichment implicating synaptic dysfunction and glial activity. In contrast, signature dimensions shared between A<sup>+</sup> and A<sup>-</sup> individuals involved language-related regions and aging-associated molecular pathways. These findings suggest that contrastive graph learning may help separate amyloid-associated functional network variation from broader background biological variability, providing insight into the heterogeneity of AD-related biomarkers and cognitive dysfunction.