Ethologovideographic Correlates of Mental Disorders in Nonverbal Behavior (Message 2: Discriminant, Thermodynamic and Network Characteristics of Facial and Pantomimic Reactions)

Background: the study of nonverbal behavior based on the ethological paradigm using computer vision technologies is currently considered as one of the approaches to objectification of mental disorders. At the same time, their analysis using multidimensional data processing methods, primarily from the standpoint of thermodynamic and network analysis, is of particular interest. Purpose of the work: to summarize the results of discriminant, thermodynamic and network analysis of facial and pantomimic activity of patients with neurotic disorders and schizophrenia spectrum pathology. Patients, Comparison Group and Methods: 19 patients with schizophrenia spectrum disorders (Sch), 23 with neurotic disorders (ND). 22 healthy subjects made up control group (CG). Analysis of non-verbal behavior was carried out using the biometric video analytics complex “MIX VR-19” based on action units (AU) of the FACS. Results: the logarithm of the thermodynamic probability of the system of facial reactions was maximal in individuals with ND (50.2), minimal in people with Sch (33.1), and it occupied an intermediate position in the CG (44.2). The system organization was also noticeably lower in Sch (0.08) than in ND (0.11) and in CG (0.14). Analysis of the graphs showed that the nodes page ranks in healthy subjects were characterized by the highest weight of AU61, AU1, AU63 and AU64. For ND, the nodes with maximum page ranks were AU14 and AU64. In Sch, the highest rank was observed for nodes AU62 and AU2. Conclusions: the greatest contribution to the differentiation between the groups of people with ND, Sch and CG was made by such AU as ironic smile, squinting and opening the mouth. The dynamics of entropy and organization of the facial-pantomimic reactions system in the CG reflected the tension of adaptation mechanisms at stages with a predominance of cognitive load; in Sch, there was an insufficiency of such tension during load, and in ND, tension was observed at stages with affective and personal themes. The graph of facial-pantomimic reactions in healthy individuals was determined by integrativeness with a large number of connections between nodes; In patients with HP, the graph was sequentially connected, while in individuals with SR it was represented by isolated clusters of AU.