Comportamento não verbal e análise de credibilidade: busca de evidências quantitativas para o auxílio na detecção de mentiras

Abstract

Nonverbal behaviors have been employed in lie detection from a diagnostic perspective, emphasizing the assessment of deception through incongruence between verbal content and emotional expression. This study aims to examine the role of nonverbal behaviors as a complementary component in the lie detection process. The first chapter presents a systematic review of studies published between 2020 and 2024 in major international databases. Following the application of inclusion criteria, ten articles were selected for analysis. The findings indicate that the identification of nonverbal behaviors in deceptive contexts requires careful consideration of contextual and situation-specific factors. The second chapter reports a statistical analysis based on data extracted from the MU3D database, which includes discourses categorized by veracity (truth vs. lie) and emotional valence (positive vs. negative). Facial movements, eye movements, and heart rate signals were extracted using remote photoplethysmography through the rPPG-Toolbox integrated with MediaPipe. The results demonstrate that nonverbal behaviors show statistical significance primarily in relation to the emotional valence of the discourse, but are insufficient for reliably distinguishing veracity.