Prediction and classification of anxiety-related psychological scale and VR sickness based on autonomic physiological responses during VR treatment in patients with social anxiety disorder

Introduction Social anxiety disorder (SAD) can accompany emotional symptoms as well as physical reactions. The assessment and real-time measurement of SAD is difficult in real-world. Objectives This study aims to predict the severity of specific anxiety states and virtual reality (VR) sickness in SAD patients by a machine learning model based on only quantitative measuring of autonomic physiologic signals during VR therapy sessions. Methods In total, 32 individuals with SAD symptoms were enrolled in VR participatory sessions. We assessed patients’ specific anxiety symptoms through Internalized Shame Scale (ISS) and Post-Event Rumination Scale (PERS), and VR sickness through Simulator Sickness Questionnaire (SSQ). Specific anxiety symptoms and VR sickness were divided into severe and non-severe states based on the total score of each scale by K-means clustering. Logistic regression, Random Forest, Naïve Bayes classifier, and Support Vector Machine were used based on the physiological signal data to predict the severity group in subdomains of ISS, PERS, and SSQ. Results Prediction performance (F1 score) for the severity of the ISS mistake anxiety subdomain was higher than other scales with 0.8421. For VR sickness, prediction performance for the severity of the physical subdomain was higher than the non-physical subdomain with 0.7692. Conclusions The study findings present that mistake anxiety and physical sickness could be predicted more accurately by only autonomic physiological signals, suggesting these features are probably associated with autonomic responses. Based on the present study results, we could provide the evidence for predicting the severity of specific anxiety or VR adverse effects only based on in-situ physiological signals. Disclosure No significant relationships.

Introduction: Grief is a normal and not necessarily pathological psychological process that occurs after the loss of a family member or loved one with its psycho-affective consequences, external manifestations and rituals.Although mourning processes can be associated with losses of different types (employment, housing, baseline situation, housing), we will refer to mourning for the loss of a loved one.For some people, social networks facilitate the expression of feelings and experiences of grief, connecting with the emotional support of other friends and loved ones.However, the presence of accounts belonging to these deceased persons, the persistence of photos and memories that periodically appear on the screen without the person being able to choose them, can make it difficult to process the mourning.Objectives: The aim of this paper is to consider the beneficial and detrimental factors of social media during a grieving reaction after the loss of a loved one.Methods: For the preparation of this work, a bibliographic review on the subject has been carried out.Likewise, the clinical information provided by patients during our evaluations has provided critical views on what has been published in this regard.
Introduction: Social anxiety disorder (SAD) can accompany emotional symptoms as well as physical reactions.The assessment and real-time measurement of SAD is difficult in real-world.
Objectives: This study aims to predict the severity of specific anxiety states and virtual reality (VR) sickness in SAD patients by a machine learning model based on only quantitative measuring of autonomic physiologic signals during VR therapy sessions.Methods: In total, 32 individuals with SAD symptoms were enrolled in VR participatory sessions.We assessed patients' specific anxiety symptoms through Internalized Shame Scale (ISS) and Post-Event Rumination Scale (PERS), and VR sickness through Simulator Sickness Questionnaire (SSQ).Specific anxiety symptoms and VR sickness were divided into severe and nonsevere states based on the total score of each scale by K-means clustering.Logistic regression, Random Forest, Naïve Bayes classifier, and Support Vector Machine were used based on the physiological signal data to predict the severity group in subdomains of ISS, PERS, and SSQ.Results: Prediction performance (F1 score) for the severity of the ISS mistake anxiety subdomain was higher than other scales with 0.8421.For VR sickness, prediction performance for the severity of the physical subdomain was higher than the non-physical subdomain with 0.7692.

Conclusions:
The study findings present that mistake anxiety and physical sickness could be predicted more accurately by only autonomic physiological signals, suggesting these features are probably associated with autonomic responses.Based on the present study results, we could provide the evidence for predicting the severity of specific anxiety or VR adverse effects only based on in-situ physiological signals.
Disclosure: No significant relationships.Keywords: physiological response; social anxiety disorder; virtual reality; Anxiety Eating Disorders EPV0689 Emotion dysregulation, dissociation and body dissatisfaction mediate the relationship between trauma exposure and ED symptoms Introduction: The current study tests the relationship between eating disorder (ED) symptoms and trauma exposure.The mechanisms via which trauma is related to ED symptoms have not been sufficiently examined.This study examines the complex role of dissociation and emotional dysregulation in the context of trauma, BMI, ED symptoms and body dissatisfaction (BD).Objectives: We hypothesized that dissociation and emotional dysregulation would mediate the relationship between trauma exposure and ED symptoms / BD.We further hypothesized that BMI would play a moderating role in this association.
European Psychiatry S577