Virtual reality (VR) has emerged as a promising tool in the field of mental health, especially for anxiety management. Unlike pharmacological treatments, VR-based interventions are non-invasive, carry fewer side effects and can be deployed at scale. Yet one major challenge persists: not everyone benefits from them. The reasons for these differences have remained largely unknown, until now.
A recent study titled Enhanced Brain-Heart Connectivity as a Precursor of Reduced State Anxiety after Therapeutic Virtual Reality Immersion, published in Advanced Science in 2025, provides new insight into why some individuals respond positively to VR anxiety therapy while others do not. Led by Idil Sezer and Anton Filipchuk, the research team examined both brain and heart activity during a VR-based anxiety intervention. Their findings offer the first evidence that connectivity between the brain and the autonomic nervous system, known as brain-heart coupling, may predict therapeutic response.
The anxiety gap: why non-pharmacological treatments don’t help everyone
Anxiety disorders affect over 300 million people worldwide [1] and the demand for accessible, non-drug interventions has surged. VR therapy has already demonstrated effectiveness in multiple studies, whether for fear of flying, acrophobia, social anxiety or general stress reduction.
Still, clinical results reveal a recurring issue: heterogeneity of response. In other words, some individuals show rapid reduction in anxiety after VR therapy, while others report little to no improvement.
Most existing research has tried to solve this problem by modifying the psychological content of interventions: adjusting narratives, exposure intensity or relaxation guidance. However, this approach assumes the limitation lies in intervention design, rather than in individual neurophysiology.
The neurovisceral integration model [2] proposes that emotional regulation depends on dynamic coordination between cortical control regions (prefrontal and cingulate regions) and autonomic nervous systems (vagal and sympathetic regulation). If these neural-autonomic interactions differ across individuals, it could explain why some people are biologically more receptive to VR-based therapies.
Until recently, this idea hadn’t been directly tested in a VR setting. The study by Sezer et al. [3] is one of the first to examine how brain and heart signals interact during VR therapy and how these interactions differ between treatment responders and non-responders.
Understanding this mechanism could transform how VR-based mental health solutions are deployed, moving from a one-size-fits-all approach to a personalised therapeutic framework.
What’s at stake: personalisation, efficiency and patient outcomes
Variability in treatment response is not a trivial problem, it directly affects clinical outcomes, adoption by healthcare providers and long-term trust in digital therapeutics. If clinicians and digital health platforms cannot anticipate who will benefit from VR, they risk wasting time, reinforcing patient frustration and delaying access to effective care.
This is why biomarkers of response matter. In pharmacotherapy, biomarkers have revolutionised cancer treatment through personalised medicine. In mental health, the equivalent progress has been slower, partly due to the complexity of neural and emotional processes.
However, Sezer et al.’s study marks a turning point by identifying physiological indicators measurable in real time, EEG beta band activity and heart rate variability (HRV) that together predict anxiety reduction.
If validated on a larger scale, these findings could enable:
– adaptive VR protocols that adjust in real time based on user physiology;
– improved clinical triage to allocate the right intervention to the right user;
– objective monitoring of therapy progress beyond self-reported anxiety scales;
– integration of biofeedback and neurofeedback to reinforce therapeutic gains.
Experimental design. Illustration of the two videos displayed in the virtual reality headset, simultaneously with EEG and ECG acquisition.
Inside the experiment: immersive VR, EEG and heart signals
The study recruited 27 healthy participants reporting mild everyday anxiety. They were exposed to a brief therapeutic VR session developed by Healthy Mind, combining a calming immersive landscape, a hypnotic script promoting mental disengagement and paced breathing guidance to modulate autonomic activity. Anxiety levels were measured using the validated State-Trait Anxiety Inventory (STAI-Y1) [4].
Physiological activity was recorded using two modalities:
1. Electroencephalography (EEG): a 17-channel system targeting midline cortical areas implicated in emotional and cognitive control, with particular attention to beta frequency bands (13-30 Hz) linked to attention, cognitive engagement and top-down regulation [5].
2. Electrocardiography (ECG): to quantify heart rate variability (HRV), especially high-frequency (HF) power associated with parasympathetic activity and low-frequency (LF) power reflecting sympathetic and parasympathetic balance [6][7].
The key methodological innovation lies in analysing directed connectivity between EEG and HRV signals using Granger causality, revealing not only whether brain and heart interact but also in which direction. This allows researchers to distinguish whether the brain influences the heart (top-down regulation) or vice versa (bottom-up interoceptive feedback).
Participants were then divided into:
– responders for those whose STAI-Y1 scores decreased;
– non-responders for those with no change or increase.
EEG beta band relative spectral power response in each subgroup.
Results of the study: the mechanistic differences between treatment responders and non-responders
The results reveal a clear and compelling difference between the two groups. Responders showed a significant increase in midline beta activity during the VR session, particularly across central and frontal electrodes.
This activation pattern, associated with sustained attention and cognitive regulation, suggests active engagement of emotional control networks. In contrast, non-responders showed minimal beta modulation, implying reduced cortical engagement.
Mechanistic explanation for treatment response‐related group differences.
Responders also displayed stronger EEG-to-HRV directional influence, particularly on HF power, indicating that brain activity effectively modulated heart autonomic responses during immersion. This dynamic brain-heart synchrony aligns with adaptive emotional regulation, whereas non-responders showed weak or absent coupling [8].
In addition, the study found that participants showing higher HF power, reflecting stronger parasympathetic modulation, tended to experience greater anxiety reduction.
Limitations to keep in mind
Like all exploratory studies, this work has limitations that must be acknowledged. First, the sample size of 27 participants restricts generalisability and replication is needed across diverse populations and clinical diagnoses, including generalised anxiety disorder or post-traumatic stress disorder (PTSD).
Second, the intervention combined several components, VR immersion, hypnotic suggestion and paced breathing, making it difficult to isolate which mechanism drives the observed brain-heart coupling.
Third, the study assessed immediate response. Future research must examine whether these biomarkers predict long-term therapeutic benefit.
Finally, while Granger causality offers valuable insight into directed interactions, combining it with additional causal inference techniques (e.g. transfer entropy or dynamic causal modelling) would strengthen mechanistic understanding.
Nevertheless, these findings represent a pivotal step toward biologically informed, personalised VR therapy.
How Neuromind can build on these findings
Like the study’s vision, Neuromind’s scientific approach focuses on understanding the biological signatures underlying mental states to improve prediction, diagnosis and treatment. Using multimodal signals (EEG, heart rate variability, behavioural data, etc.), Neuromind develops explainable AI models that map the interplay between neural activity and emotional regulation.
Building on these results, Neuromind can extend the use of biomarker-informed triage: before starting a full VR or digital therapy programme, short assessments could detect neurophysiological profiles similar to “responders”. This would allow clinicians and digital platforms to adapt intervention strategies, such as integrating targeted neurofeedback or stress-regulation training for individuals showing low brain-heart synchrony.
Neuromind invests in adaptive closed-loop systems, where real-time physiological monitoring dynamically adjusts the therapeutic environment. Such systems could, for instance, adapt immersion intensity or breathing guidance based on instantaneous feedback from EEG and HRV. This paradigm transforms digital therapy from static to responsive.
By integrating these findings, Neuromind contributes to the emergence of technology for mental health, bridging neuroscience and empathy to make VR anxiety therapy more adaptive, measurable and human-centered. To learn more about our work, don’t hesitate to contact us.
[1] World Health Organization, Anxiety disorders, 2025.
[2] Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61(3), 201–216.
[3] Sezer, I., Filipchuk, A., Moreau, P., et al. (2025). Enhanced brain-heart connectivity as a precursor of reduced state anxiety after therapeutic virtual reality immersion. Advanced Science, 12(3), 2304021.
[4] Spielberger, C. D. (1983). State-Trait Anxiety Inventory (STAI) Manual. Consulting Psychologists Press.
[5] Engel, A. K., & Fries, P. (2010). Beta-band oscillations-signalling the status quo? Current Opinion in Neurobiology, 20(2), 156–165.
[6] F. Shaffer, R. McCraty, C. L. Zerr, Front. Psychol. 2023, 5, 1040.
[7] Quintana DS. Statistical considerations for reporting and planning heart rate variability case-control studies. Psychophysiology. 2017 Mar;54(3):344-349.
[8] Critchley HD, Eccles J, Garfinkel SN. Interaction between cognition, emotion, and the autonomic nervous system. Handb Clin Neurol. 2013;117:59-77.