EDUCATION

My thesis focused on the computational and neural mechanisms that support novel human interactions, how we communicate and make sense of each other’s intentions without relying on shared language. Using a mix of behavioural experiments, computational modelling, eye-tracking, and EEG, we employed controlled tasks, Tacit communication game, to study how people signal and interpret intentions through actions alone in novel situations.

In the first study, we found that when people lack access to a shared language, they rely on universal knowledge grounded in the shared physical world as prior expectations, and then intentionally deviate from those expectations to produce strategic signals that help their partners infer their goals. We developed a computational model of this behaviour and validated it across two independent behavioural datasets, alongside model-based EEG and pupillometry analysis. 

In the second study, we asked whether such expectation-based signalling might offer a simpler alternative to traditional Theory of Mind (ToM) mechanisms in novel communication contexts. We compared the performance of our model against ToM-based agents in a human–computer interaction setup, and found that expectation-based strategies matched more complex mentalizing while also replicating the human behaviour better. 

Additionally, as part of a collaborative project (CML) with research teams in China, I contributed to studies on the sense of agency, the feeling of control over one’s actions and their consequences. This involved both empirical experiments and a meta-analysis of the literature, and resulted in two co-authored papers exploring intentional binding effects and the neural network of sensory attenuation. 

For my PhD dissertation, click here. 

During my Master’s I focused on understanding the brain networks involved in motor control, particularly the role of cortical and subcortical regions like the premotor cortex, supplementary motor area (SMA), putamen, and thalamus.

My thesis explored a long-standing question in motor neuroscience: Do the basal ganglia initiate voluntary movement, or do they only support and execute movements generated by the cortex? To address this, I conducted fMRI data collection and applied Bayesian Mixed Linear Non-Gaussian Acyclic Models (BMLiNGAM) to examine effective connectivity patterns across different computational states of movement.

We compared two distinct phases of motor processing:

• The early phase, associated with movement planning and origination, revealed a top-down network structure, with strong influence from the cortex, particularly PMd and SMA.

• The later phase, related to movement control and execution, showed a shift toward bottom-up feedback loops, implicating subcortical structures like the putamen and thalamus.

This project gave me a strong foundation in network analysis, Bayesian modelling, and fMRI data interpretation, while deepening my interest in the dynamic interplay between cortical and subcortical systems in complex behaviour.

For my Bachelor’s thesis,  I investigated how people weigh costs and benefits when deciding whether to seek or avoid cognitive closure, the desire for a firm answer rather than uncertainty. I designed and conducted a survey-based study, collected data from participants, and performed statistical analyses to examine how motivational factors influence this decision-making process. This early project sparked my long-standing interest in the cognitive mechanisms behind human judgment and decision-making.