Research on psychotherapeutic processes aims to clarify the mechanisms of change, refine theoretical models, and offer valuable tools for clinical practice. A central focus is on microprocesses, brief yet significant moments (emotional breakthroughs, alliance ruptures, and repairs). These phenomena have been studied primarily through transcriptions, video coding, and detailed observations: valuable methods, but complex, time-consuming, and sometimes influenced by the evaluators' subjectivity. From early observations of "micromovements," facial expressions have emerged as key indicators: expressive variations signal engagement, patient-therapist synchronization strengthens the alliance, and specific nonverbal cues anticipate turning points. Recently, automated tools have been developed to detect basic emotions and Action Units (AUs) with high temporal precision, offering more objective and scalable measures. The study ims to determine whether automated analysis of AUs reliably identifies clinically relevant moments, particularly impulses, and to evaluate its usefulness for both research and clinical practice.