Modelling of Spatiotemporal EEG and ERP Brain Data for Dynamic Pattern Recognition and Brain State Prediction using Spiking Neural Networks: Methods and Applications in Psychology
This thesis aimed to improve modelling and understanding of spatiotemporal brain data underpinning human behaviour with the use of a brain-inspired artificial intelligence technique: spiking neural networks (SNN). These networks incorporate both space and time information of brain data into one unifying model, allowing to capture spatiotemporal relationships and the trajectory of sequentially activated brain areas in response to different types of stimuli under different mental states. SNN models were used in this thesis for both dynamic pattern recognition and pattern prediction in two real-life empirical scenarios from neuroinformatics: (1) neuromarketing and (2) mindfulness training. In neuromarketing study, results showed that how early marketing materials are perceived at an unconscious level of information processing and elucidated the underpinning dynamics of these processes. In mindfulness study, SNN could capture changes in brain data in relation to mindfulness intervention across individuals with different levels of depression. This thesis expands the field of cognitive science and artificial intelligence on three main empirical contributions: as a generic framework for dynamic pattern recognition and prediction in brain data; as a model of consumers’ behaviour for detection and prediction of preference; and as a mental wellbeing application for detection and prediction of brain responses to an intervention.