Brain mapping, biomarker identification and using machine learning method for diagnosis of anxiety during emotional face in preschool children.

BACKGROUND

Due to the importance and the consequences of anxiety, the goals of the current study are brain mapping, biomarker identification and the use of an assessment method for diagnosis of anxiety during emotional face in preschool children.

METHOD

45 preschool children participated in this study. Functional Magnetic Resonance Imaging (fMRI) data were taken in fearful and angry conditions. The functional connectivity (FC) for the limbic system were extracted by ROI-to-ROI method. The fMRI biomarkers (FC) were given to machine learning models as input features to diagnose anxiety in children for angry and fearful conditions.

RESULT

The results of the brain mapping comparisons between anxiety and the non-anxiety showed that there was an increased FC between medial prefrontal cortex (MPFC) and right lateral amygdala (RLA) and a decreased FC between left anterior hippocampus (LAH) and left posterior hippocampus (LPH) in the angry condition. There was an increased FC between the pairs of regions, RLA- right anterior hippocampus (RAH), MPFC-LPH, and RAH-LPH in fearful condition. It is possible to use the FC between LAH- right medial amygdala (RMA) and the FC between left medial amygdala (LMA)-RMA, LMA-RLA, LMA-RAH, and left lateral amygdala (LLA)-RLA instead of IQ in angry and fearful conditions, respectively. Based on metrics such as accuracy, recall, precision, and area under the receiver operating characteristic curve, the Logistic Lasso Regression model outperformed the other model in diagnosing anxiety.

CONCLUSION

With these findings, psychiatrists and psychologists can have a better understanding of the brain connectivity in children.