Trends in the use of affective computing in e-learning environments

Considering that emotions have a great impact on motivation, reasoning, and decision making, affective computing methods, that were designed to attempt to understand and respond to human emotional states, have been used in more than one field including e-learning. Thus, a systematic literature review was conducted on 4 search engines resulting in a set of papers that were filtered in a systematic way until we obtained a corpus of 27 papers. Data were extracted to answer four research questions concerning the use and efficacy of affective computing in e-learning in recent years. We found out that the majority of studies about emotion recognition use uni-modal systems in which facial expressions emotion detection is the most present. The major research purpose is designing/building systems, approaches, methods, detectors for emotion recognition. For the e-learning environments, the most present is conversational agents. The emotions detected or used are basic emotions, non-basic emotions, learning-centered emotions, trait emotions, or a combination of two or three of them. This systematic literature review also provides the major findings, challenges, and future research.