Recognition Of Social Anxiety Based On Spectral Statistics Of RR Intervals In ECG

Social anxiety is a discomfort or a fear when a person is in a real or imaginary social interaction that involves a concern of being inevitably judged or evaluated by others. It is a meaningfol research for individuals’physical and mental health to effectively identify the status and degree of social anxiety. The fluctuation characteristics of RR intervals of 30 subjects in public speaking states and its intrinsic mechanism with social anxiety are investigated in this thesis.Public speaking anxiety is one of the most representative subtypes of social anxiety. First of all, an experimental group and two control groups are set up to speak in four states of public speaking experiments, namely, the baseline state, speech ready state, no-audience speaking state and public speaking state where electrocardiograph (ECG) signals are obtained. Then the peaks of R waves of ECG signals are regarded as a fractal stochastic point process on a time-line{ti}. The RR intervals are formed through the automatic positioning and manual editing of R peaks. The standardized RR intervals are formed through the utilization of the mathematical model to magnify heartbeat spectrum{t1} that averages neighboring spacing of RR intervals. Meanwhile, RR intervals can be analyzed according to their nearest neighboring spacing, a kind of measure method based on the level fluctuations of energy spectra. The Brody distribution, an interpolation distribution of two extreme models which are named Poisson and Wigner distribution, is introduced to fit the nearest neighboring spacing of RR interval. The energy level repulsion parameters α of the nearest neighboring spacing is determined through controlling the fitting residual sum of squares and fitting precision. Finally, the intrinsic mechanism between a and social anxiety is investigated through the comparison of the range characteristics of a under the four states of the experimental group and two control groups.Results reveal that the repulsion parameter a differs significantly between the baseline state and the speech ready state, but varies lightly between two consecutive baseline states. Taking the baseline and speech ready states as an experimental group and the two consecutive baselines as a control group, it can be shown that the significant differences are caused by social anxiety. The repulsion parameter a differs significantly between the public speaking state and the baseline state as well as the states of no-audience speaking and public speaking. Besides, a high recognition rate of 95% can be achieved when a is taken as a feature in the binary classification on the existence of social anxiety. And the Gini index of a is the highest one compared with other features. It can be summarized that the energy level repulsion parameter a can be employed as a valid index to identify social anxiety. Further more, the fact that the degree of social anxiety of each state is quantified by the subjects’ self-evaluation scores helps to draw the final conclusion that the degree of social anxiety can be judged by the numerical value of repulsion parameter a.