In vivo amygdala nuclei volumes in schizophrenia and bipolar disorders

AbstractBackgroundAbnormalities in amygdala volume are well-established in schizophrenia and commonly reported in bipolar disorders. However, the specificity of volumetric differences in individual amygdala nuclei is largely unknown.MethodsPatients with schizophrenia disorders (SCZ, n=452, including schizophrenia, schizoaffective and other psychotic disorders, mean age 30.7±9.2 (SD), females 44.4%), bipolar disorders (BP, n=316, including bipolar I and II, 33.7±11.4, 58.5%) and healthy controls (n=753, 34.1±9.1, 40.9%) underwent T1-weighted magnetic resonance imaging. Total amygdala and nuclei volumes as well as intracranial volume (ICV) were estimated with Freesurfer (v6.0.0). Analysis of covariance and multiple linear regression models, adjusting for age, age2, ICV and sex, were fitted to examine diagnostic group and subgroup differences in volume, respectively.ResultsBilateral total amygdala and all nuclei volumes, except the medial and central nuclei, were significantly smaller in patients relative to controls. The largest effect sizes were found for the basal nucleus, accessory basal nucleus and cortico-amygdaloid transition area (partial η2 > 0.02). The diagnostic subgroup analysis showed that reductions in amygdala nuclei volume were most widespread in schizophrenia, with the lateral, cortical, paralaminar and central nuclei being solely reduced in this disorder. The right accessory basal nucleus was marginally smaller in SCZ relative to BP (t = 2.32, p = 0.05).ConclusionsOur study is the first to demonstrate distinct patterns of amygdala nuclei volume reductions in a well-powered sample of patients with schizophrenia and bipolar disorders. Volume differences in the basolateral complex (lateral, basal, accessory basal nuclei) may be putative neuroimaging markers for differentiating schizophrenia and bipolar disorders.