In times of growing importance and emphasis on improving academic outcomes for young people, their academic selves/lives are increasingly becoming more central to their understanding of their own wellbeing. How they experience and perceive their academic successes or failures, can influence their perceived self-efficacy and eventual academic achievement. To this end, 'cognitive emotions', elicited to acquire or develop new skills/knowledges, can play a crucial role as they indicate the state or the "flow" of a student's emotions, when facing challenging tasks. Within innovative teaching models, measuring the affective components of learning have been mainly based on self-reports and scales which have neglected the real-time detection of emotions, through for example, recording or measuring facial expressions. The aim of the present study is to test the reliability of an ad hoc software trained to detect and classify cognitive emotions from facial expressions across two different environments, namely a video-lecture and a chat with teacher, and to explore cognitive emotions in relation to academic e-self efficacy and academic adjustment. To pursue these goals, we used video-recordings of ten psychology students from an online university engaging in online learning tasks, and employed software to automatically detect eleven cognitive emotions. Preliminary results support and extend prior studies, illustrating how exploring cognitive emotions in real time can inform the development and success of academic e-learning interventions aimed at monitoring and promoting students' well being.