Electrophysiological systems underlying human domain-general cognition and selective attention

Abstract

Understanding how neural dynamics support domain-general cognition and selective attention is crucial for uncovering the mechanisms that enable flexible cognitive processing across tasks. This thesis investigates the contributions of electrophysiological systems—including oscillatory, aperiodic, and evoked responses—to these processes, using a combination of cutting-edge techniques such as MEG/EEG, MEG-fMRI fusion, and concurrent TMS-EEG. The research aims to examine how these electrophysiological responses encode task demands across diverse cognitive domains and guide attentional control. In Chapter 2, I use MEG/EEG to investigate the contributions of aperiodic and oscillatory systems to human domain-general cognition. By applying irregular-resampling auto-spectral analysis (IRASA) and multivariate pattern analysis (MVPA), I show that both aperiodic (broadband power, slope, and intercept) and oscillatory (theta, alpha, and beta power) components encode task demand and content across three cognitive tasks. Aperiodic broadband power, in particular, shows the strongest decoding performance and cross-task generalisability, highlighting its role in domain-general cognition. In Chapter 3, I extend these findings with an fMRI study using the same task sets to examine the relationship between aperiodic broadband power and the domain-general multiple-demand (MD) network. Using representational similarity analysis (RSA)-based MEG-fMRI fusion, I demonstrate that aperiodic broadband power most strongly shares task demand-related representational structure with MD activation, establishing a link between this electrophysiological signal and the MD network as a basis for domain-general cognition. In Chapter 4, I focus on the role of alpha oscillations and event-related potentials (ERPs) in selective attention. I first analyse the temporal dynamics of alpha oscillations and ERPs in encoding attentional information using EEG. Then, I apply rhythmic-TMS (rh-TMS) at individual alpha frequency over the right intraparietal sulcus to causally manipulate parietal alpha power and ERPs. Concurrent EEG reveals that, compared to arrhythmic-TMS, alpha rh-TMS selectively enhances the decoding of spatial attentional information (i.e., where to attend) while leaving other attentional information (i.e., what to attend to or visual feature information) unaffected. These findings demonstrate how evoked and oscillatory activity encode complementary aspects of selective attention and reveal a critical role of IPS-mediated evoked and oscillatory activity in guiding spatial attention. In Chapter 5, I further examine whether the role of alpha activity in spatial attention is purely oscillatory or also involves aperiodic components. Analysing three open EEG datasets, I show that both oscillatory and aperiodic alpha activity contribute to spatial attention encoding and lateralisation effects. Re-analysing the TMS-EEG dataset from Chapter 4, I find that rh-TMS enhances oscillatory alpha power while reducing aperiodic alpha power relative to arrhythmic TMS. Despite these opposing effects, rh-TMS improves spatial attention representation in both oscillatory and aperiodic alpha signals, with each predicting different aspects of behavioural performance. These findings suggest a dual mechanism of alpha activity in spatial attention that oscillatory and aperiodic alpha activity may play distinct but complementary roles. In conclusion, this thesis provides a comprehensive account of how distinct electrophysiological responses support human domain-general cognition and selective attention. Aperiodic broadband power emerges as a robust neural correlate of cognitive demand and closely aligns with the representational structure of the MD network. Oscillatory and aperiodic alpha activity, along with ERPs, contribute to the encoding of complementary aspects of attentional information, with causal evidence from TMS-EEG demonstrating their specific involvement in spatial attention. Together, these findings advance our understanding of how diverse electrophysiological systems support flexible, goal-directed cognition.