Hang-Yee Chan 陳恆義

When people watch the same video, their brains don't respond identically — but the more they do, the better that video tends to perform in the real market. This is the core insight behind neuroforecasting: the synchrony of neural responses across a small sample of people carries predictive signal about what a much larger population will prefer, buy, or remember. We identified the temporal lobe and cerebellum as key substrates of this effect across two fMRI studies using TV commercials and movie trailers, and extended the framework to show that shared affective patterns from brief picture-viewing episodes can decode emotional responses to entirely different, naturalistic content.

The natural next step was to ask whether the knowledge embedded in neuroimaging data could be transferred into scalable tools that don't require a scanner. NeuroViT does exactly that: a vision transformer trained on consumer neural responses learns brain-aligned image representations that predict how product images drive decisions — turning a relatively small neuroimaging investment into a broadly applicable model for image evaluation.

Output: NeuroImage (2019); NeuroImage (2020); NeuroViT preprint (2025)

Ask someone what a brand means to them and they'll give you an answer — but that answer is a pale summary of a rich, multisensory, largely non-verbal set of associations stored in the brain. The neural profiling work takes a different route: instead of asking, we compare brain responses while participants passively view social scenario photographs and while they actively imagine a brand. The overlap between the two produces an individual neural profile of brand image that outperforms self-report in predicting co-branding suitability and tracks population-level brand strength rated by an independent sample.

The ad liking work asks a related question: when an ad works, what is actually happening in the viewer's brain, and when? Drawing on three fMRI datasets spanning 113 participants, 85 video ads, and two countries, we decode the psychological processes — emotion, social cognition, memory, perception — that unfold second by second during ad exposure. Emotion kicks in first, social cognition sustains, and it is the social-affective signal that best predicts whether a population will like the ad — a finding that holds up out-of-sample and beats both self-report and conventional anatomical neuroimaging approaches.

Output: Journal of Marketing Research (2018); Journal of Marketing Research (2023)

Before a piece of information travels through a social network, it has to pass through a brain. An fMRI study on the gatekeeper role reveals that simply thinking about sharing — rather than reading for oneself — changes how information is neurally encoded. Encoding consistency in the default mode network drops when participants consider broadcasting to their social feed, and the effect is strongest for those who occupy high-betweenness positions in their real-life networks. Gatekeeping isn't just a decision; it reshapes perception.

The question of what makes a message shareable turns out to have a remarkably consistent neural answer across cultures. In a cross-cultural fMRI study, activity in regions tied to self-relevance, mentalizing, and valuation predicted individual sharing intentions — and those neural predictors transferred to new cultural contexts without retraining. Large-scale behavioral work in 10,000+ participants confirms the mechanism: people share content they find relevant to themselves or their social ties, across topics from health to politics. A mega-analysis of 16 independent neuroimaging studies then stress-tests this picture at scale, finding that brain responses — particularly in self-referential and affective regions — reliably explain message effectiveness across labs, stimuli, and populations. A parallel strand examines what happens when persuasive messages don't arrive in isolation: in real social media feeds, health messages compete with contradictory content, and that competition meaningfully shapes their impact.

Output: Cerebral Cortex (2021); JEP: General (2023); PNAS (2023); PNAS Nexus (2025); PNAS Nexus (2025); Social Science & Medicine (2025)

Thinking about yourself and thinking about others feel like very different mental acts, yet they activate overlapping territory in the brain — particularly medial prefrontal cortex. The question is whether that overlap reflects genuinely shared computation or just spatial proximity. Using multivoxel pattern classification, we show the two can be pulled apart: the neural signatures of self- and other-related thought are separable even where they anatomically overlap, with implications for theories of self-other distinction and mentalizing.

Two further collaborative projects address social behavior from different angles. A preregistered direct replication of the dietary self-control literature puts the vmPFC–striatum connectivity finding under rigorous scrutiny — asking whether the neurobiological marker of successful regulation holds up when the methodological chips are down. And a developmental neuroimaging study tracks how conformity to peer influence is implemented in the brain from adolescence into early adulthood, finding that the neural systems engaged shift with development in ways that map onto changing susceptibility to social pressure.

Output: Cerebral Cortex (2022); Human Brain Mapping (2022); Developmental Cognitive Neuroscience (2023)