Orbitofrontal cortex and amygdala as key regions for the classification of small animal phobia: A structural MRI machine learning study.

Abstract

Small animal phobia (SAP) is an anxiety disorder characterized by an intense fear triggered by small animals. Existing studies on SAP have primarily used univariate analyses and small, unbalanced samples, leading to inconsistent findings. Many brain structures have been proposed as potential candidates for encoding SAP, including the amygdala, the orbitofrontal cortex, the insula, the cingulate gyrus, the hippocampus, the precuneus, and the putamen. However, no previous study has systematically compared the predictive role of each of these regions to determine which one provides the most accurate classification of individuals with SAP. To address this question, we tested the predictive power of these brain regions associated with SAP. Structural MRI images of 32 individuals with SAP and 90 matched healthy controls were analyzed by using multiple Binary Support Vector Machines (BSVM). Our results showed that the most critical regions for classifying SAP, in order of significance, were the orbitofrontal cortex (balanced accuracy = 80.83%), the amygdala (76.15%), the insula (71.35%), and the posterior cingulate (71.25%). However, only the orbitofrontal cortex and the amygdala consistently exceeded chance levels across all additional parameters considered (precision, specificity, sensitivity, and F1-score). In conclusion, this study enhances the understanding of the neural basis of small animal phobia, suggesting new research directions and diagnostic strategies using advanced machine learning methods applied to neuroimaging data. Clinical implications are also discussed.