Date of Award

4-25-2026

Thesis Type

PhD

Document Type

Thesis

Divisions

Institute for Advanced Studies

Institution

Universiti Malaya

Abstract

This study integrates conversational analysis principles and neural activity to investigate natural cognition and examine event-related brain changes. A wireless electroencephalography (EEG) device was used in an ecological validity setting to analyse pre-stimulus brain activity preceding conversational turns. This study aims to identify aspects of the conversation that reflect its naturalness in the aspect of listening and the corresponding brain activity by analysing the signals before and after a turn in a conversation. This research contributes to an ecologically valid understanding of natural conversation as a social action that enables seamless communication. Seven Malaysian participants (aged 19–22, M=20.46) engaged in natural conversations while their brain activity was recorded via a 32-electrode wireless-EEG device using the 10-20 placement standard. The participants were matched in ethnicity, education, and English proficiency. Nine images with varying emotional valence served as stimuli embedded within the conversations to evoke ecologically valid cognitive processing. EEG data underwent preprocessing, Fast Fourier Transform (FFT), power spectral analysis, Independent Component Analysis (ICA), and Multiple Artefact Rejection Algorithm (MARA) filtering. Two key methodological innovations were developed to troubleshoot and correct temporal inaccuracies inherent in wireless-EEG recordings, and preserve millisecond-level accuracy for time-locked analyses of naturally occurring events. Results revealed all 14-aspects of natural conversation to be present, as well as topic initiation, topic progression, turn durations, turn distribution, and total conversation time to reflect natural variability similar to real-world conversations. The N100 and P300, which are typically measured in an experimental settings, were observed in a real-time, real-person, and real-conversation design of this study. Anticipatory alpha desynchronisation was localised predominantly in frontal brain regions, previously linked to cognitive processing, rather than parietal or occipital areas. Additionally, higher emotional intensity elicited relatively greater alpha power in the -2000 to -1000 ms prestimulus window. Despite being at pattern identification level, these findings suggest that pre-stimulus neural activity during conversation extends beyond memory and general anticipation and may involve aspects of neurobiological processing not observed before. The study’s findings have implications for human-machine interfaces as well as a framework for developing neural markers for clinical disorders.

Initial

khm

Additional Information

Thesis (PhD) – Institute for Advanced Studies, Universiti Malaya, 2026.

Download

Share

COinS