Neurofeedback-driven Emotional Regulation Training in a virtual reality environment: a machine learning approach using OpenBCI

This paper addresses the design and development of an advanced neurofeedback system for training in emotional regulation skills and competencies; the system integrates a Virtual Reality (VR) platform with a 16-channel OpenBCI device for real-time capture of electroencephalographic (EEG) signals. The main objective of the research lies in the application of machine learning algorithms, specifically Random Forest and K-Nearest Neighbors (KNN), for the classification of emotional states in terms of valence and arousal. These algorithms achieve an accuracy of up to 83% for arousal classification and 90% for valence. EEG signals are processed and classified in real time and the results are integrated into a virtual reality environment created in Unity. This adaptive environment changes according to the detected emotional states, allowing for more precise regulation. In addition, a diaphragmatic breathing protocol has been developed within the virtual reality environment as an intervention strategy for emotional regulation. The system is in its final stage of piloting to establish the efficacy of the system.