JMIR Publications

ABSTRACT

Background:

Phobias are unfortunately common in the United States, interfering with daily living by dictating physical, mental, and emotional well-being. Successful treatment of phobias can lead to improved quality of life, greater self-efficacy, and reduced anxiety. Virtual reality provides a safe, controlled environment for exposure treatment, while artificial intelligence enables real-time adaptation of the therapeutic experience. Together, these technologies enable real-time, adaptable scenarios for safe, effective, and individualized treatment.

Objective:

This scoping review paper studies how the intersection of virtual reality (VR) and artificial intelligence (AI) can improve treatment for phobias.

Methods:

This study was conducted using IEEE Xplore and Medline databases and yielded a final sample of 14 papers from 2010 to 2014 after screening and applying eligibility criteria.

Results:

Research has shown that the most common health condition studied is acrophobia (5 of 14 examined studies or 35.71 %), with AI inputs primarily consisting of physiological data (for all examined studies) and AI outputs focused on predicting affective states—such as fear levels in 5 out of 14 total studies (35.71%), as well as electroencephalogram (EEG) classifications, severity groups, cortical states, stress scores, social anxiety disorder (SAD), and arousal states in other papers—mostly through classification or prediction methods, while few studies directly modified the VR environment.

Conclusions:

The studied papers effectively use real-time data adaptations to personalize treatments, where future work could focus on more interactive AI agents and broader clinical studies; however, this review is limited by a small sample size, restricted database usage, and limited inclusion of varied terminology in the search string. Overall, the combination of AI and VR offers a promising new approach to phobia treatment, with the potential to improve access and outcomes of care.