# Identifying Neurobehavioral Biomarkers of Anxiety and Treatment Response Using Virtual Reality, Electroencephalography, Magnetic Resonance Imaging, and Related Multimodal Assessments: A Longitudinal Study Protocol

## TL;DR

This study uses virtual reality, brain scans, and other tools to find biological signs of anxiety and how well people respond to treatment.

## Contribution

The study introduces a multidimensional, longitudinal approach to identify biomarkers of anxiety and treatment response using multimodal assessments.

## Key findings

- Multimodal data will be used to build machine-learning models predicting treatment response.
- Longitudinal analyses will track symptom changes and neural mechanisms over time.
- Comparisons will be made between responders, non-responders, and healthy controls.

## Abstract

Background/Objectives: Anxiety disorders are highly prevalent and impairing psychiatric conditions. Conventional diagnostic approaches based on symptom checklists lack biological specificity and often fail to guide treatment decisions effectively. This study protocol outlines a multidimensional, prospective investigation designed to identify behavioral and neurobiological biomarkers predictive of treatment response in individuals with anxiety-related symptoms, grounded in the Research Domain Criteria framework. Methods: This observational, longitudinal study (NCT06773585) will include a transdiagnostic sample of clinical anxiety group alongside a healthy control group (185 participants, including 145 patients with anxiety disorders and 40 healthy controls). Participants will undergo comprehensive baseline assessments, including clinical interviews, self-report questionnaires, a virtual reality (VR)-based behavioral task, electroencephalography (EEG), electrocardiography (ECG), and structural and functional brain magnetic resonance imaging. Follow-up assessments will be conducted at 2, 6, and 12 months, with recruitment and data collection planned from 2024 to 2029. These complementary modalities are integrated to capture behavioral, physiological, and neural indicators of anxiety and its treatment response. Multimodal baseline features will be used to construct machine-learning models predicting treatment response, defined as ≥40% reduction in anxiety severity scores. Longitudinal analyses will examine symptom trajectories and neural mechanisms associated with response. Neurobiological comparisons will be made across timepoints and between responders, non-responders, and healthy controls. Conclusions: By identifying objective, biologically grounded markers of anxiety and treatment response, our findings will contribute to the development of personalized assessment tools and scalable digital interventions for psychiatric care.