Not the Pearls but the Thread: Using fNIRS Hyperscanning to Investigate Real-Life Social Interactions

Hyperscanning refers to the simultaneous recording of brain activity from two or more individuals. Since its introduction in 2002, this approach has represented a major methodological advance in social neuroscience, enabling researchers to directly investigate the neural mechanisms underlying active social exchange. Around the early 2010s, the use of functional near-infrared spectroscopy (fNIRS) became increasingly popular in hyperscanning research. This trend was driven by the technique’s high tolerance for head and body movements, its noninvasive nature, and its portability, which together made it particularly suitable for examining neural dynamics during naturalistic social interactions. Over time, the hyperscanning literature has evolved along two main and interrelated trajectories. The first is methodological, focusing on improving the acquisition and analysis of brain signals in experimental settings that allow greater freedom of movement and more ecologically valid social interactions. The second is conceptual, aiming to elucidate how the brain supports active social engagement at both the intra- and interpersonal levels. In line with these directions, the present thesis is based on five experimental studies encompassing both methodological and conceptual investigations. On one side, Chapter 2 presents two experimental studies aimed at improving the identification and correction of motion artifacts in fNIRS data. The first study introduces and demonstrates the feasibility of a computer vision approach for extracting head movement information from video recordings. Using these ground-truth movement features, the second study examines the relationship between specific movement characteristics and the corresponding noise patterns in fNIRS signals. On the other side, the thesis includes three additional experimental studies designed to examine how relational factors influence patterns of interpersonal neural synchrony and functional connectivity. Specifically, Chapter 3 presents studies investigating how the linguistic features and emotional content of dialogues are associated with, and supported by, patterns of prefrontal interpersonal synchrony and within-brain functional connectivity. Chapter 4 reports a large-scale hyperscanning study involving 284 participants, which examines how interpersonal closeness and task interactivity modulate interpersonal neural synchrony. Methodologically, this thesis demonstrates the feasibility of using computer vision techniques to extract reliable movement features during fNIRS acquisition and to link these features to motion-induced artifacts. These results provide an empirical foundation for developing more robust, movement-informed pre-processing pipelines. Conceptually, the findings show that emotional content and linguistic complexity modulate both interpersonal and intra-personal neural coupling, suggesting that cognitive and affective dimensions of communication jointly shape the coordination of neural activity between and within interacting individuals. Moreover, interpersonal closeness and task interactivity were found to differentially influence synchrony patterns across prefrontal and temporo-parietal regions. Overall, this work highlights the interdependence between methodological innovation and conceptual understanding in social neuroscience. By presenting works on both computer vision–based movement annotations and naturalistic hyperscanning paradigms, the thesis contributes to a more ecologically grounded framework for studying interpersonal neural synchrony. These insights hold both theoretical significance for models of social cognition and practical potential for applications in educational, clinical, and everyday interactive contexts.