Neuroimmune State Configurations at the Amyloid-Tau Interface in Alzheimer's Disease: A Testable Framework

Neuroimmune mechanisms are central to Alzheimer's disease (AD), but a single continuum of "more inflammation" does not explain why similar amyloid-β (Aβ) and tau burdens are associated with different clinical trajectories, tissue responses, and treatment risks. We present a testable neuroimmune-state framework that integrates, rather than replaces, the AT(N) system, cellular-phase models, and transcriptional descriptions of disease-associated microglia and reactive astrocytes. The framework treats neuroimmune states as multidimensional, regionally heterogeneous configurations defined by pathological substrate, cellular program, biological target, vascular and aging context, and measurable patient-level readouts. We distinguish four meanings of the amyloid-tau interface: a biomarker interface, a spatial tissue interface, a cellular-state interface, and a clinical-progression interface. We then examine the hypothesis of phagocytic target switching, in which clearance pathways that support plaque compaction or debris removal may, under selected tau- and complement-rich conditions, contribute to removal of stressed but potentially functional synapses. We explicitly separate evidence from longitudinal human studies, cross-sectional biomarkers, human genetics, postmortem tissue, animal models, and in vitro systems. Current biomarkers cannot yet assign patients to validated discrete neuroimmune states; proposed state readers should therefore be probabilistic, allow regional coexistence and parallel vascular-immune trajectories, and be tested for incremental value beyond established AD biomarkers and co-pathologies. Finally, we organize immune-directed interventions by clinical evidence level and frame stage matching as a research strategy rather than a treatment algorithm. The model is useful only if its predictions can be prospectively supported or falsified.