Dr Manikandan V M, Associate Professor, Department of Computer Science and Engineering along with his B.Tech students authored a research study on Immersive VR Puzzle System With EEG-Based Real-Time Stress Analysis For Emotional And Cognitive Assessment with Application Number: 202541131948. Their patent is filed under the Indian Patent Office.
Abstract:
This project is created with the mindset to support people having troubles due to anxiety and stress. Which is why we came up with a solution that integrates Virtual Reality (VR) gaming therapy platform with electroencephalography (EEG). According to this pipeline, users engage in VR puzzle based scenarios that include certain tasks which require focus and concentration with gradually increasing difficulty levels. While the user is solving these puzzle tasks, there will be an EEG headband that monitors factors like stress, anxiety, focus with the help of the signals across the frequency bands which are alpha, beta, and theta, which run in sync with specific gameplay events that gives insights about their cognitive load and how well they behave in a little to medium stress induced environment. At the same time, performance(in-game) like error rate, completion time etc. are also recorded while the game is running. The important factor about this project is the combination of an immersive VR environment and a neuro-feedback analysis, unlike the traditional studies regarding EEG which records the user’s data in a static environment, or in the form of questionnaire which gives little to none insights about their actual performance in a stressful/anxious situation. Our system provides continuous, quantitative feedback and does not rely on self-reported or post-session assessment about their experience. Applications are plenty which include a stress and anxiety assessment tool in both research and clinical perspective that offers diagnostic support for therapists and psychologists by ensuring multiple sessions in this and have a track of all of the emotional states across each of the sessions. This system is developed using publicly available VR headset and EEG headband. The puzzle were designed in the Unity Engine and the neuro signals are captured using the SDK provided by the EEG headband company
Explanation in Layperson’s Terms:
Our brains emit waves for every emotion we are feeling like stress, anxiety, happiness etc. We are capturing these waves by combining puzzle games in VR with a brainwave-reading headband (EEG). The user wears a VR headset and an EEG headband and plays 3 different puzzles developed by us that slowly become harder and more confusing as you play along. Apart from fun, these puzzles are meant to gently create mental load and mild stress, in a controlled and safe way so that the user becomes more and more engrossed trying to play better.
While the person is playing, we record their brain activity, especially the strength of brainwaves in different frequency bands like alpha (usually higher when relaxed), beta (often higher during focus or stress), and theta (linked to concentration and cognitive effort).
We then capture this stream of data using time: for every second of gameplay, we know what was happening in the game and what the brain was doing. Using this data, we apply machine learning models to learn patterns that separate calmer moments from more stressed moments. The important difference from traditional studies is that we don’t just ask people afterwards how they felt or make them do static screen-based tasks. Instead, we provide an immersive VR experience and continuously measure their brain and behaviour as they handle mild stress in real time. This lays the foundation for tools that could, in future, help therapists monitor, track, and eventually adapt VR therapy based on the user’s actual mental state rather than only self-reported feelings.
Implementations:
This invention provides an approach to stress and anxiety analysis by combining immersive VR puzzle based tasks with real-time EEG monitoring. By integrating physiological data with behavioral performance metrics, the system enables accurate and objective assessment of emotional responses under increasing difficulty levels.
Application as a Stress and Anxiety Assessment Tool
Primary Use: The invention is designed to evaluate how individuals respond to gradually increasing cognitive challenge and mental pressure. Users engage in VR puzzle tasks with rising difficulty levels while their EEG signals and gameplay performance are analyzed to detect emotional strain, anxiety levels, and cognitive overload.
Advantages: Provides objective measurement of stress responses rather than relying on subjective questionnaires. Identifies stress thresholds and behavioral patterns during complex tasks. Useful for early detection of stress sensitivity or anxiety-related tendencies.
Diagnostic and Analytical Support for Mental Health Professionals
Use for Therapists: The system offers a detailed physiological and behavioral profile of a user’s emotional state during cognitive tasks. EEG data when paired with error rates, reaction times, will provide clinicians a scientific basis for understanding stress reactivity.
Advantages: Enhances clinical evaluation with quantifiable neurophysiological data.
Helps track patients’ stress or anxiety progression over time. Supports evidence-based decisions regarding therapy approaches and cognitive-behavioral interventions. Moves beyond self-reporting by providing real-time insight into emotional regulation and cognitive strain.
Longitudinal Tracking Through EEG-Based Emotional Metrics
Application in Monitoring Progress: The system stores EEG and gameplay data from multiple sessions to detect long-term trends, enabling comprehensive tracking of a user’s resilience, stress tolerance, or cognitive performance.
Advantages: Offers measurable evidence of emotional improvement or deterioration. Helps clinicians, researchers, or individuals assess the effectiveness of stress-management techniques or therapy. Enables comparison across sessions, giving a complete picture of growth or changing stress patterns.
Social Implications:
If used responsibly, the system has the potential to make mental health assessment more objective and data-informed. Instead of relying solely on self-reported stress scores, therapists and researchers gain access to continuous physiological and behavioural measurements during controlled, repeatable tasks. This can improve understanding of how different individuals respond to stress, help identify patterns that are not obvious from conversation alone, and provide more concrete evidence of progress or deterioration over time.
At the same time, there are important ethical and social considerations. EEG and behavioural data are highly sensitive and must be handled with strict privacy and security measures; storage, sharing and analysis should follow clear consent and data protection norms. The system is not a diagnostic device on its own and should not be used to “label” people without professional interpretation. There is also a risk of misuse if such technology were applied in non-clinical contexts (for example, workplaces) to monitor stress without genuine support or consent.
Collaborations:
Our setup and workspace was in the Next Tech Lab(Vikram Sarabhai Block, 203). As the Board members of NTL, Daksh & Anjana used all the computational resources from the lab without which we would not have been able to develop the puzzles and even set up the whole pipeline needed to test the integration of EEG headband and the VR headset. Some of the lab members even volunteered to help in testing the pilot development by participating to play the puzzles. The lab was equipped with a Meta Quest 3 headset(VR headset) and as for the EEG headband, we had received a Seed grant worth 50,000 INR in March 2025 and procured the same in November of the same year. A huge thanks goes to the Seed grant committee and the Dean Research office. We even thank Dr. Ayesha Haroon, Head of Department of Psychology, as her insights about people suffering with stress and anxiety and existing tools about diagnosing them, helped us develop our puzzles.
Future Research Plans:
As a next step, this work will be extended into a formal pilot study with a larger and diverse participant group. We plan to call around 100–150 participants to use our VR–EEG system and complete the puzzle-based tasks. We will record their brainwave activity. This expanded dataset will allow us to perform an accurate statistical analysis, train and validate our machine learning models on a wider population, and study individual differences in stress responses more systematically. Based on the results of this pilot study, we aim to publish a journal paper documenting the full pipeline, experimental protocol, model performance and key findings. The larger-scale data will also help us refine our feature extraction methods, improve classification accuracy for different stress levels, and explore personalization of models for individual users. In the longer term, this foundation can be used to investigate real-time adaptation of VR environments based on detected mental states and to design more targeted therapeutic scenarios for anxiety and stress-related conditions