Seeing And Communicating With Your Mind: A Novel EEG Signal Based Interactive Visual Meditation System

Abstract

In recent years, mental health regulation has gained attention. Approaches are generally external, involving professional treatment, or internal, focusing on self-regulation like meditation. Internal methods, particularly meditation, are effective for managing stress and improving well-being. However, meditation involves complex techniques and is often challenging for beginners to perform correctly. This study aims to develop a technique, termed Visual Meditation, to help beginners engage effectively in meditation. Using Electroencephalography (EEG), we monitor participants’ brain waves during meditation to provide real-time feedback. Specific brain wave patterns, such as alpha or beta waves, indicate the onset of deep meditation. Based on these EEG readings, dynamic visual scenarios are created to guide the user’s mental state, enabling even untrained individuals to meditate correctly. We gathered EEG data from over 100 testers with varied meditation experience and demographics to create a dynamic meditation visualization. SVM was employed for wave classification and meditation state prediction. For user interaction, we utilized VQGAN-CLIP for scene adaptation and element transformation, linking the meditation environment with EEG data. This technique aims to democratize effective meditation practices for mental health. After we have tested and trained a large number of samples, the model enabled real-time mental state interaction, accelerating entry into meditation by about 30%. The technique provides a foundation for future research in meditation and neurophysiology.