Proposal for an interlocutor model indicating therapeutic effectiveness in clinical interviewing

Résumé En

Within the Franco-German Mephesto project, various actions were implemented over the course of 3 years. In particular, the Semagram team at Inria Grand-Est was involved on two levels. The first was data collection at the Pasteur Hospital in Nice, involving 11 subjects suffering from schizophrenia and 13 subjects at high risk of developing psychosis, i.e. prodromal patients at the Centre de la Madeleine. The data collection consisted of a clinical interview in the presence of a double eye-tracking system, enabling measurement of the subject's attention via the average of intra- and inter-regional ocular saccades of the face, for example.In parallel, a multimodal cross-sectional study with a prospective longitudinal perspective was carried out, involving the development of a meaningful interlocutory model of therapeutic efficacy. It is based on the transcripts of 17 subjects with schizophrenia and 17 control subjects. Based on discourse analysis, the aim is to uncover the complex, including the necessary conditions and other associated phenomenon(s), which is akin to a so-called "clinical/therapeutic" effect.Methodologically speaking, the idea is to discern a cognitive-psycholinguistic profile of the situation in which a speaker finds himself and can benefit from the action of his interlocutor, which would be comparable to a "clinical/therapeutic" effect.In the very short term, the idea is to test this model on the data collected in Mephesto. This confrontation of the model with Mephesto data will also involve testing the hypothesis that specific gaze movements by the psychologist can play a similar or complementary role to backchannels in the support activity.We plan to test models of diagnostic support and follow-up that rely on other elements such as voice intonation or physiological cues, whose complementarity we wish to evaluate.