Normative Modeling of Amygdala Subfields and Disease-Specific Vulnerabilities Across Neurological Disorders

As an integrative center of the limbic system, the amygdala and its subfields participate in diverse neural functions and demonstrate heterogeneous structural changes in various neurological disorders. In this study, three-dimensional T1-weighted MRI data from 22,258 healthy participants were analyzed using normative modeling to establish normative references for the volumes of the amygdala and its subfields. These models elucidated the developmental and aging curves of the amygdala and identified distinct ages at peak volume across subfields, ranging from 24.6 to 43.7 years. Based on these references, individual centile scores were calculated for 4132 patients with mild cognitive impairment, Alzheimer's disease, Parkinson's disease, cerebral small vessel disease, multiple sclerosis, and neuromyelitis optica spectrum disorder. These centile scores not only captured disease-specific deviation patterns of the amygdala subfields but also demonstrated utility in disease classification. Moreover, the centile scores correlated with clinical measures of cognitive function, emotional state, and motor performance across disorders. These findings suggest that normative references of the amygdala subfields provide insights into cross-disease alteration patterns and offer potential neuroimaging biomarkers to aid differential diagnosis and clinical assessment.