AI-enabled multimodal neuroimaging for neurotransmitter mapping in normal aging and age-related disease

Aging reshapes neurotransmitter systems through nonlinear and region-specific shifts that alter excitatory-inhibitory balance, weaken metabolic coupling, and destabilize large-scale neural networks. PET, MRS, molecular MRI, and optical imaging quantify elements of these trajectories, but modality-specific artifacts, cross-site variability, and limited spatial or temporal resolution fragment mechanistic interpretation. Artificial intelligence now enables the integration of these heterogeneous signals into harmonized, multimodal representations that link molecular alterations to circuit-level dynamics and clinical outcomes. This review demonstrates how AI-enabled fusion can transform neurotransmitter imaging by correcting acquisition bias, enhancing reproducibility, and revealing hidden dependencies among GABAergic, glutamatergic, cholinergic, dopaminergic, and serotonergic systems. We advance three hypotheses: that conserved neurotransmitter coupling patterns distinguish healthy from pathological aging; that excitatory-inhibitory reorganization reflects compensatory signaling rather than linear degeneration; and that AI-driven multimodal integration can generate mechanistic predictors of cognitive, affective, and motor decline. By embedding PET-, MRS-, and MRI-derived features into explainable architectures, AI models can identify early neurochemical inflection points, stratify individuals by neurotransmitter vulnerability, and support precision diagnostics. AI-enabled multimodal neuroimaging, therefore, establishes a foundation for mechanistic neuroscience in aging, reframing neurochemical decline as a dynamic interplay between adaptation and vulnerability rather than a uniform trajectory of loss.