A Web application for characterizing spontaneous emotions using long EEG recording sessions

Emotions are important in daily life and in several research fields, especially in Brain Computer Interface (BCI) and Affective Computing. Usually, emotions are studied by analyzing the brain activity of a subject, monitored by Electroencephalography (EEG), functional Magnetic Resonance Imaging (fMRI) or functional Near Infrared Spectroscopy (fNIRS), after some external stimulation. This approach could lead to characterization inaccuracies, due to the secondary activations produced by the artificial elicitation and to the subjective emotional response. In this work, we design a web application to support spontaneous emotions characterization. It is based on a database for EEG signals where a large amount of data from long recording sessions, collected from subjects during their daily life, are stored. In this way, EEG signals can be explored to characterize different spontaneous emotional states felt by several people. The application is also designed to extract features of specific emotions, and to compare different emotional states. Researchers all over the world could share both raw data and classification results. Since large datasets are treated, the application is based on strategies commonly used in big data managing. In particular, a column-oriented database is used to store a huge amount of raw EEG signals, while a relational database is employed to keep metadata information. A web application interface allows the user to communicate with the repository and a computational module performs the features extraction.