Title : The NERVE framework: AI-personalized virtual reality and closed-loop neuromodulation for trauma-linked symptom states
Abstract:
Psychiatry still relies heavily on retrospective symptom reports and static diagnostic categories, despite the fact that many clinically relevant symptoms emerge as dynamic, context-dependent brain states. This creates a major translational challenge: how can psychiatry move from observing symptoms after they occur toward experimentally evoking, detecting and adaptively regulating them in real time?
Post-traumatic stress disorder (PTSD) provides a compelling translational model for neuroadaptive psychiatry because clinically meaningful symptom states — including hyperarousal, threat reactivity, intrusive memory, affective dysregulation and avoidance — can be ethically and experimentally evoked, measured and potentially modulated. At the same time, advances in AI-driven virtual reality exposure therapy (VRET), multimodal biomarkers and closed-loop neuromodulation suggest the possibility of more dynamic and personalized circuit-based interventions.
This presentation introduces the NERVE Framework (Neuroadaptive Evoked Regulation via Virtual Environments), a translational neuropsychiatric architecture designed to integrate personalized symptom provocation, multimodal state estimation and adaptive neuromodulatory intervention within a closed-loop framework. The central aim is not to propose a finalized therapeutic protocol, but to define a clinically grounded and experimentally testable framework for transitioning from symptom provocation to real-time adaptive regulation.
The framework was developed through an extensive translational synthesis of recent literature across virtual reality exposure therapy, computational psychiatry, digital phenotyping, behavioral and physiological biomarkers, EEG-informed personalization, inflammatory correlates of stress dysregulation and closed-loop neuromodulation technologies including TMS and DBS, combined with clinically informed observations from trauma-focused psychological practice. The proposed model is organized as a provocation–detection–intervention loop: AI-personalized VR environments evoke trauma-relevant symptom states; multimodal physiological, behavioural and affective signals estimate state dynamics in real time and neuromodulatory parameters are adaptively adjusted according to timing, circuitry and individual neurophysiological profiles.
Current evidence supports the feasibility of several components independently: VR can evoke ecologically valid trauma-related responses, multimodal AI systems can infer affective and physiological state changes, and adaptive neuromodulation research highlights the importance of timing, state-dependence and individualized targeting. However, these advances remain largely siloed across disciplines. The major translational gap is the absence of unified neuroadaptive systems capable of integrating personalized symptom provocation, real-time state detection and adaptive circuit-level intervention within a single closed-loop architecture.
In conclusion, PTSD may provide an ideal translational testbed for neuroadaptive closed-loop psychiatry by enabling experimentally accessible yet clinically meaningful symptom-state modulation. The NERVE Framework reframes precision psychiatry from “the right treatment for the right diagnosis” toward “the right signal, in the right patient, at the right moment, for the right symptom state.”
