Joint analysis of verbal and nonverbal interactions in collaborative E-learning

We present here some preliminary analysis of audio-video interactions involving two groups of learners performing collaborative learning tasks. Living in two different countries, their mental representations are different and produce what we called a clash of contexts ("socio-cognitive misunderstanding"). The verbal dimension of the video was annotated by experts. The nonverbal, affective, dimension was tagged automatically. We found an interesting correlation between verbal and affective channels. Particularly, on both side, a kind of "Eureka effect" was detected as a key moment when learners understand each other and when their emotion changes, from frustration to happiness.