The State-Trait Anxiety Inventory (STAI) and the Hamilton scale for anxiety (HARS) are two of the most important scales employed in clinical and psychological realms for the evaluation of anxiety. Although the reliability and sensibility of these scales are widely demonstrated there is an open debate on what exactly their scores reflect. Neuroimaging provides the potential to validate the quality and reliability of clinical scales through the identification of specific biomarkers. For this reason, we evaluated the neural correlates of these two scales in a large cohort of healthy individuals using structural neuroimaging methods.
Neuroimaging analysis included thickness/volume estimation of cortical and subcortical limbic structures, which were regressed on anxiety inventory scores with age and gender used for assessing discriminant validity. A total of 121 healthy subjects were evaluated. Despite the two anxiety scales, at a behavioral level, displaying significant correlations among them (HARS with STAI-state (r = 0.24; P = 0.006) and HARS with STAI-trait (r = 0.42; P < 0.001)), multivariate neuroimaging analyses demonstrated that anatomical variability in the anterior cingulate cortex was the best predictor of the HARS scores (all β's ≥ 0.31 and P's ≤ 0.01), whereas STAI-related measures did not show any significant relationship with regions of limbic circuits, but their scores were predicted by gender (all β's ≥ 0.23 and P's ≤ 0.02).
Although the purpose of HARS and STAI is to quantify the degree and characteristics of anxiety-like behaviors, our neuroimaging data indicated that these scales are neurobiologically different, confirming that their scores might reflect different aspects of anxiety: the HARS is more related to subclinical expression of anxiety disorders, whereas the STAI captures sub-dimensions of personality linked to anxiety.