Medical simulations allow medical trainees to work within teams to develop their self-regulated learning (SRL) and socially shared regulated learning (SSRL) skills. These skills are imperative in optimizing performance and teamwork and could be reflected in physiological responses given by learners. This study examines how medical trainees' self-regulatory patterns can predict their psychophysiological responses, specifically their electrodermal activity (EDA), by employing supervised machine learning (ML). Sixty-two (N = 62) medical residents at a Canadian university participated in this study. Participants were grouped into 19 teams, with each completing one medical simulation with an appointed "leader" and "team members." Simulations were part of medical residents' curriculum and used high-fidelity manikins capable of mimicking physiological activity as "patients." Audio-video recordings of each simulation were coded for (1) behaviors (posture and gestures, facial expressions, and vocalics) and (2) regulation strategies, including SRL and SSRL, derived and adapted from the literature to fit a medical context. Psychophysiological measurement of EDA was collected using "Empatica E4" bracelets throughout the simulations. Raters coded the regulatory interactions between the "leader" and "team member" at the "beginning," "escalation," and "peak" of each simulation. Results indicated that varying SRL and SSRL codes could predict EDA based on the regulatory needs of learners at different segments of the simulation. These findings contribute to the literature on applying ML modeling to predict psychophysiological responses of learners and to furthering our understanding of the use of predictive modeling within multimodal data in naturalistic learning environments.