University of Twente Student Theses


Finding neural correlates for social relationships using EEG hyperscanning

Lenz, Dominik (2017) Finding neural correlates for social relationships using EEG hyperscanning.

[img] PDF
Abstract:Brain-computer Interfaces are a growing field of Human-computer Interaction that is gaining importance with sensors becoming cheaper and more accessible. This paves the way to these devices being used in daily life. A possible field of application is affective computing, in which the computer attempts to be environmentally aware of the human aspects of the user, such as emotions and the social context. This project explores the ability for BCI (in this case EEG hyperscanning) to be used to detect the social context, by attempting to find neural correlations for social relationships between two users in a joint-attention setting. The experiment consists of two users who can belong to either of two dyad classes (”strangers” or ”lovers”) getting exposed to a series of visual stimuli while their brain-activity is being recorded (EEG recording using 2 BioSemi Active2). The metric that is investigated primarily is the Inter-brain weighted phase lag index (WPLI) as defined by [Vinck et al., 2011]. The results of this experiment, based on a user test with 6 dyads, show a weak significant difference (.8 CI) between the dyad groups in the alpha and theta frequency range. The conclusions drawn are that there are clear indications for the WPLI being a usable metric for detection of social relationships, however the joint-attention task used in this experiment is a rather passive form of interaction, while other experiments with more active tasks seemed to cause stronger differences in the signals. This might hint at the effect primarily stemming from active interaction.
Item Type:Essay (Master)
1989, Nederland
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Interaction Technology MSc (60030)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page