Autonomic nervous pattern analysis of trait anxiety

Abstract

High trait anxiety (HTA) causes cognitive control deficit, implicit emotion regulation deficit and risks of emotional disorders. The current work designed a paradigm for autonomic nervous pattern analysis of trait anxiety. The paradigm includes three phases: (1) physiological data acquisition of autonomic nervous activity in resting baseline status, during the preparation and presentation of an evaluated speech, and in recovery status after stressful task; (2) feature extraction and selection of the physiological data; (3) pattern recognition of high/low trait anxiety through commonly used classifiers. By the operation of the above paradigm, we analyzed 99 samples of respiratory (Resp) data and RR interval series calculated from the ECG data of 99 subjects. We have found that: (1) the autonomic nervous activity of trait anxiety has obvious gender difference; (2) After eliminating gender difference by modeling of trait anxiety for each gender, the HTA and low trait anxiety (LTA) populations can be distinguished by the decision tree classifier with accuracies of 66.67% and 100% for females and males, respectively; (3) for the females, the critical two-dimension feature subset of HTA vs. LTA classification comes from the resting baseline status, stress recall status and stress withdrawal status; (4) for the males, the critical two-dimension feature subset of HTA vs. LTA classification comes from the resting baseline status and the stress withdrawal status. The results have shown that HTA and LTA have baseline autonomic nervous difference, and the stress reactions of autonomic nervous system of HTA and LTA populations are different for different genders.