Investigating the effects of virtual patients’ nonsensical responses on users’ facial expressions in mental health training scenarios

This report investigates how clinician-participants react to virtual patients’ sensical vs. nonsensical responses in a training simulation that aims to help clinicians acquire empathetic skills toward high-risk patients with symptoms of the Suicide Crisis Syndrome (SCS). Two suicidal virtual patients were developed, and clinician-participants interactions with them were recorded. Their facial emotions were analysed in three key moments: after a baseline sensical response, after a nonsensical response, and after the following sensical response. We compared their basic facial emotions aggregated into Negative and Positive facial affective behaviors (FABs). We describe our study involving ten clinician-participants and the results of the facial expression analysis with Noldus FaceReader. Our results suggest that nonsensical responses from virtual humans have an overall impact in both Positive and Negative facial affective emotions, and may lead to an increased percent of time participants demonstrate Negative facial affective behaviors when interacting with virtual humans. We discuss several aspects regarding the impacts and importance of considering nonsensical responses in the context of virtual human-based interactions.