TY  - THES
AB  - This thesis brings together two formerly distinct topics: The research about neural foundations of interaction and machine-mediated interaction, particularly using BMIs. Both have been active fields of research for years and there are countless contributions on these topics. To my knowledge, however, studying machine-mediated interaction settings as a special case of human interaction is new to neuro interaction research. In particular the use of BMIs, which are the most direct connection between humans and machines, as a technique in neuro interaction research is unprecedented.

This thesis aims to pave a way for this new approach. The first experiment (HExMInE) aims to verify the general feasibility of the approach, in particular to evaluate whether or not neural correlates of interaction (in particular hyper-connectivity) still occur when interaction is machine-mediated. The second experiment (iCusss) then is intended to showcase the potential of this approach in a fully featured (BMI) machine-mediated interaction experiment.

The thorough exploration of this approach’s potential is undoubtedly by far too ambitious for a single thesis. The same holds for the (thorough) investigation of the impact of machine-mediation on human interaction and its neural correlates. As this research advances, though, the results will undoubtedly shed light on important aspects of human interaction and human-machine interaction and can be expected to have major impact on the design of future machine-mediation technology. Therefore, this new approach addresses highly relevant research goals.

Besides these rather general aspects, both experiments address specific research questions, relevant for neuro interaction research: In the HExMInE experiment I compare neural connectivity during interaction and during solo action of one participant. In the iCusss experiment I compare neural connectivity during cooperation with independent, concurrent action. The main method for this evaluation is hyper-scanning and -analysis aiming for neural connectivity – within a participant and across participants. This connectivity analysis is conducted on different frequency bands, addressing several of the
standard neural rhythms in human EEG and thereby contributing to the interpretation of their roles. These roles have been another active research topic for years and still evidences for new facets regarding their interpretation/function is being accumulated.

Additionally, different scientifically relevant topics are addressed as a side-effect when pursuing my main research goals: For the HExMInE experiment a new type of training for the mental strategy of Motor Imagery, often employed in BMIs, is tested. For the two experiments two different robots which are diametrical in many aspects are employed. In particular one is highly anthropomorphic while the other is the exact opposite. Up to now, there are relatively few publications on BMI usage of more than one participant simultaneously. This thesis contributes indirectly to the field of multi-user BMIs by demonstrating its feasibility even for very complex settings. In the course of the PhD project the development of the UBiCI BMI software framework was advanced. And finally, I employ two different BAPs for the BMI control, both of which are correlated with some function vital for interaction (P300 ↔ attention and ERD ↔ motor co-representation) with the intention to allow space for interesting side effects to occur in the neural recordings.

This thesis deviates from two different, well explored paths of research at once, converging in and pioneering a brand new direction of neuro interaction research. I hope this path will lead research to the neural foundations of human interaction from a new, different angle, allowing an illumination of new aspects of what Schilbach et al called “the dark matter of social neuroscience”.
DA  - 2020
DO  - 10.4119/unibi/2946075
LA  - eng
PY  - 2020
TI  - More Than One - Observing Human Brains in Uncommon Interaction Scenarios
UR  - https://nbn-resolving.org/urn:nbn:de:0070-pub-29460759
Y2  - 2024-11-22T04:04:44
ER  -