Dissertation: Electroencephalogram-Based Attention Detection: Integrating Event-Related Potentials and Machine Learning Methods
Yuan Qin’s doctoral dissertation, Electroencephalogram-Based Attention Detection: Integrating Event-Related Potentials and Machine Learning Methods, systematically investigates the neural mechanisms by which attention guides selective processing in complex environments, and develops more effective EEG-based methods for attention detection.
The dissertation comprises four parts: the first two focus on experimental psychology, revealing how attention responses and dynamic patterns vary during static regularity detection and simulated driving tasks. The latter two address methodological advances, introducing support vector machine–based approaches combined with novel threshold postprocessing and channel reordering techniques to enhance the accuracy of predicting action preparation and on-going states while reducing computational demands. Furthermore, the study examines spatial normalization across nine attention-related paradigms, demonstrating its significant benefits for decoding performance in between-subjects settings.
Findings show that EEG reliably detects potential differences induced by diverse stimuli in both static and dynamic contexts. Notably, even subtle stimuli or anticipatory attentional states can be effectively captured with appropriate processing methods. By highlighting the role of threshold postprocessing, channel reordering, and spatial normalization, this dissertation demonstrates the broad potential of EEG-based approaches for attention research and future applications in cognitive monitoring
M.Sc. Yuan Qin defends their doctoral dissertation in Cognitive science "Electroencephalogram-Based Attention Detection: Integrating
Event-Related Potentials and Machine Learning Methods". Opponent is Professor Benjamin Cowley (University of Helsinki) ja Associate Professor Tuomo Kujala. The language of the event is English.
The dissertation event can be attended in L304 (The Liikunta building).