Investigation of social anxiety of patients with schizophrenia using virtual avatar

Backgrounds: Patients with schizophrenia show a tendency to avoid social interaction because they feel great anxiety during their social interactions. As a result, they have trouble in making social relationships. This trouble is largely due to the fact that these patients show emotional withdrawal as well as passive/apathetic social withdrawal. Social deficits of these patients have been overcome through social skills training programs, which provide them opportunities to experience various social situations by role play method. Objectives: This study was conducted in order to investigate whether an interaction with a virtual avatar can evoke patient's social anxiety and the relationship between patient's symptom severity and social anxiety traits. Method/Tools: Fifteen patients with schizophrenia and fifteen controls were recruited. A male and a female avatar were generated. 2x2x3 (group, avatar's gender and avatar's emotional expression) was used in this experiment. Each avatar has three emotional expressions: are happy, neutral and angry. Subjects performed an introduction task in six conditions (gender x emotional expression) in random order. The task was composed of "approach", "listening to avatar's introduction" and "introduce oneself to the avatar". After all six tasks were performed, subjects completed a State-Trait Anxiety Inventory (STAI) questionnaire. In addition, patients' symptom severity was evaluated using the Positive and Negative Syndrome Scale (PANSS). Results: The social anxiety level to avatar's emotional representation was significantly different in the two group (p<. 001). In the control group, the level was lowest when subjects coped with an avatar expressing "happy," while the highest anxiety level was shown when they faced an avatar expressing "angry." However, in the patient group, the level difference was not significant between "neutral" and "happy" avatar. Patients only showed a significantly high anxiety level when they experience the "angry" avatar condition compared to other two conditions. In a correlation analysis between patient's anxiety level and their symptom severity, social anxiety for "happy" and " neutral" avatars was positively correlated with the negative syndrome of PANSS (happy : r=.539, p=.038, neutral : r=.533, p=.041). Particularly, the anxiety level for "happy" and " neutral" avatars was positively correlated with two subscales (N1: blunted affect, N4: passive/apathetic social withdrawal) of the negative syndrome of PANSS (N1. happy: r=.549, p=.034/neutral: r=.536, p=.039) (N4. happy: r=.536, p=.039/neutral: r=.658, p=.008). Conclusion: Through this study, we discovered that virtual avatar displaying emotional expression can cause social anxiety in patients with schizophrenia and that a patient's symptom severity was correlated with social anxiety level. It could be said, from these results, that the more severe the negative symptoms a patient has, the greater social anxiety they will feel. This might be due to their passive, apathetic and isolated trait. They may tend to recognize socially positive emotion as fearful stimuli. This inference could be supported by the relationship between anxiety and subscales of negative symptom of PANSS. In this sense, we could say that a virtual avatar could provide an opportunity for patients to experience emotionally induced social situation. In addition, it could be used for training patients to cope effectively by experiencing emotions close to reality as well as to find out the clinical characteristics related to patient's symptoms.