Research Focus
Research in our lab aims to understand how goals, motivations and emotions affect how people perceive and respond to their physical and social environments. We probe this question across multiple levels of cognitive processing, from perception and attention, to memory and decision-making, to affective and social reasoning. Our work is thus situated at the intersection of cognitive neuroscience, affective science and social psychology. At present, projects in the lab can be broadly divided into two categories. The first uses psychophysics, computational models and functional brain imaging (e.g., fMRI and fNIRS) to study affective influences on learning and decision-making. This line of work uses computational models (e.g., drift diffusion models, Bayesian cognitive models and reinforcement learning models) to decompose affective biases in cognition into separable psychological compliments, which can then be mapped onto the brain to identify the specific computational role performed by different brain regions. The second uses naturalistic stimuli (e.g., stories, and movies), natural language processing models (e.g., word or contextual embedding models), and functional brain imaging (e.g., fMRI and fNIRS) to study how goals and motivations bias the processing of narratives in naturalistic “real-world” settings.
Methodological Approaches
The lab uses a wide variety of methodological approaches from cognitive neuroscience, affective science, social psychology and computer science. Key techniques that are used heavily in the lab at the moment include: fMRI, computational modeling, natural language processing, deep learning networks, fNIRS, eye-tracking, and psychophysiology.
Affective Influences on Learning and Decision-Making
Emotion and affect guide our attention, which in turn shapes how we learn and make decisions. Our research had previously found that people learn to attend to aspects of the environment relevant for predicting reward via mechanisms of reinforcement learning, and attention in turn constrains learning and value signals in the brain to guide adaptive decision-making (Leong, Radulescu et al., 2017; Niv et al., 2015). In other work, we have also investigated how social value (e.g., perceived trustworthiness or competence) biases neural and computational processes involved in person perception and social decision-making (Leong & Zaki, 2018; Morelli, Leong et al., 2018).
More recently, our lab has been investigating how affective biases in visual attention give rise to suboptimal perceptual decisions. For example, we find that people are biased towards seeing percepts they are motivated to see, a phenomenon commonly referred to as “wishful seeing”. Using a combination of fMRI and computational modeling, we demonstrate that motivational biases in perceptual decision-making can be decomposed into perceptual and response components that are associated with distinct neural processes (Leong et al., 2019). We have also showed that the perceptual component of motivational biases are associated with pupil-related arousal processes, suggesting that they are mediated by the locus coeruleus norepinephrine (LC-NE) neuromodulatory system (Leong et al., 2021).
Current lab projects:
What are the neurocomputational mechanisms underlying affective biases in learning and decision-making (Calabro et al., 2023)?
What is the role of the locus coeruleus norepinephrine (LC-NE) neuromodulatory system in mediating affective biases in learning and decision-making?
How are affective biases in learning and decision-making impacted in psychopathology (Rossi-Goldthorpe et al., 2021)?
Do approach and avoidance motivations have distinct effects on learning and decision-making?
Neuroscience of Narratives
Narratives have a powerful hold over the human mind. People are more often convinced by a compelling story than by concrete facts. Furthermore, people use narratives to organize their thoughts and communicate ideas. Our past research has explored the neuroscience of how narratives are remembered and communicated (Chen, Leong et al., 2017; Zadbood et al., 2017). For example, we found that viewing, listening to an audio description, and verbally recalling events from a movie elicited similar event-specific neural patterns that were common across participants. These shared responses across brains are thought to reflect shared representations of narrative content. In addition to these empirical studies, we have also contributed to a large open fMRI dataset of participants listening to audio stories (Nastase et al., 2021).
More recently, we used fMRI and semantic analyses of real-world political content to study the psychological and neural processes that drive partisan biases in how people interpret political messages (Leong et al., 2020). We show that neural responses diverge between conservatives and liberals watching the same political video, and that this divergence was strongest during moments in the video with threat-related and moral-emotional language. In ongoing work, we are building NLP models that will allow us to predict when and how political messages are interpreted by people with different political beliefs.
Current lab projects:
Can we predict narrative engagement and affective experience from patterns of functional connectivity (Jin & Leong, 2022A, 2022B)?
How do individual differences in personality affect how people interpret ambiguous narratives?
How do prior beliefs bias how people remember and retell narratives?
What psychological and neural processes drive the divergent processing of political information?
