Introducing e-Motions: a novel intraoperative test for social cognition mapping. Triple validation in normative, schizophrenia, and autism spectrum disorder populations

Abstract

Background: Social cognition is essential for daily functioning, as it influences quality of life, return to work, and interpersonal communication. While schizophrenia and autism spectrum disorder (ASD) have been the paradigmatic conditions in which social cognition is markedly impaired, emerging evidence suggests that up to 30% patients with brain tumors may experience persistent deficits in this domain. Despite its clinical relevance, social cognition remains insufficiently studied in neuro-oncology and lacks dedicated intraoperative assessment tools specifically tailored for its use during awake brain surgery.

Method: This study introduces e-Motions, a new test designed ad hoc to address this gap. The e-Motions test comprises 34 four-second video stimuli depicting two hyper-realistic avatars (one male, one female) expressing complex emotions. These avatars were developed using an AI-based facial motion capture system applied to 60 professional actors. Validation was performed in three groups: (1) healthy adults (n = 226), (2) individuals with schizophrenia (n = 33), and (3) ASD (n = 30).

Results: Internal consistency (KR-20), test-retest reliability (ICC), and correlations with established social cognition tools (Reading the Mind in the Eyes [RMET], Ekman-60 faces test [Ekman-60F], and the Movie for the Assessment of Social Cognition [MASC]) were evaluated. The e-Motions test demonstrated high global internal consistency (KR-20 = 0.86) and good test-retest reliability (ICC₂,₁ = 0.73). Scores showed positive moderate correlations with both lower-level mentalizing tests (RME: ρ = 0.44; Ekman-60F: ρ = 0.48) and higher-level mentalizing test (MASC: ρ = 0.57). Discriminative power was strong for distinguishing healthy participants from individuals with schizophrenia (AUC = 0.89) and ASD (AUC = 0.79).

Conclusions: e-Motions test is the first AI-based test created ad hoc for awake brain mapping. Its good internal consistency and significant correlation with low-level and high-level mentalizing tests make this tool a novel, ecological and promising way to identify critical regions involved in social cognition during awake brain mapping. Future studies should be performed to demonstrate its validity preserving this complex higher-order cognitive construct during and after brain tumor surgery.