Building a virtual psychological counselor by integrating EEG emotion detection with large-scale NLP models

Abstract

Depression is a common psychological disorder, and the combination of psychological help and medication is the main approach to treat depression. However, traditional psychological counseling requires professional therapists to spend much time with patients, and the demand for therapists exceeds the supply. Therefore, this paper proposes a virtual psychotherapist architecture: combining EEG emotion detection and NLP large models to detect the patient's emotions in real time and adjust AI communication strategies to provide psychological counseling to patients. This paper introduces EEG emotion detection from several aspects, such as emotion models, EEG feature extraction, databases, emotion elicitation methods, and classification models. It also reviews the development of NLP large models in recent years. Then, it presents the theoretical methods for constructing a virtual psychotherapist using EEG emotion detection combined with NLP large models: the direct suggestion method and the prompt method. Finally, it points out the current problems and future development directions of virtual psychotherapists.