D'Mello's research is in the fields of affective and attentional computing, multimodal interaction, speech and discourse processing, and intelligent learning environments. His team conducts basic research on affective and cognitive states (e.g., confusion, boredom, mind wandering) across a range of interaction contexts, develops real-time computational models of these states, and integrates the models in affect- and attention- aware intelligent technologies. Their research uses a range of techniques such as eye tracking, speech recognition, physiological sensing, computer vision, nonlinear time series analyses, discourse modeling, and machine learning. Interaction contexts include intelligent learning environments, educational games, collaborative problem solving, classroom discourse, text, scene, and film comprehension. Data is collected in the lab, online, and in schools.