Brain Representations of Unique Emotional Experiences

This project explored the complexities of emotional experiences and their neural basis. It aimed to overcome the challenge of eliciting and measuring a wide range of emotions in a controlled setting. By using emotionally evocative autobiographical videos, we could elicit a broad spectrum of emotions in participants and measure their individual emotional trajectories using a collaborative filtering model.

The study highlighted significant variability in emotional responses and their correlation with changes in the brain's default mode network. Unlike traditional experimental paradigms using decontextualized stimuli, we found emotional experiences varied considerably across participants. Our findings revealed that 90% of the variance in participants’ ratings could be explained by an average of 7 dimensions, a departure from the typical 2-5 dimensions identified in prior work.

Using a probabilistic Shared Response Model, we identified common latent emotional responses that were fairly reliable across participants. However, there was dramatic variability in how individual emotional experiences were transformed into these common responses. Although participants reported differing experiences when watching the videos, we found a shared neural response to emotion changes, especially in the Default Mode Network. This research opens new avenues for understanding how emotions influence memory and communication, emphasizing the deeply personal nature of emotional experiences.



A Brief Intervention to Motivate Empathy in Middle School

Authors: Erika Weisz, Patricia Chen, Desmond C. Ong, Ryan W. Carlson, Marissa D. Clark, Jamil Zaki

Source: Journal of Experimental Psychology: General

Empathy tracks socioemotional adjustment during early adolescence, yet adolescents this age tend to show reductions in empathy compared with younger children. Here we took a novel approach to building empathy among early adolescents in four middle schools (n = 857). Rather than addressing the ability to empathize, we targeted the motivation to empathize. To do so, we leveraged strategies demonstrated to change motivation among early adolescents: social norms and mindsets. Compared with those in other conditions, students who received a norms-based intervention reported greater motivation to empathize with others, which was in turn associated with increased peer-reported prosocial behaviors, as well as lower levels of loneliness and aggression. The effects of this norms condition were strongest at schools with relatively high engagement with the intervention. Findings suggest a novel avenue for increasing empathy among early adolescents—focusing on peer-driven motivation—and underscore the importance of context in shaping intervention outcomes.

Surprise Signals Changing Affective Experiences in Naturalistic Sports Spectating

Authors: Marissa D. Clark, Luke J. Chang

Source: Neuron

Why do we enjoy watching games? Athletic competitions have a rich history in western civilization dating back at least to ancient Greece, and spectating sports and other types of competitions has become a popular international pastime. Watching games provides a sense of collective identity and can foster communal pride when one’s preferred team defeats a rival, leading to changes in mood, production of hormones, and even an increase in aggressive behavior. Spectatorship provides a common set of knowledge and shared experiences that can facilitate conversations and social connection between individuals. However, beyond these social aspects, considerably less is known about the psychological and neural processes involved in watching a game.

Organizing Academic Papers Using Ward Clustering

As a second-year student facing my qualifying exams, I was tasked with organizing a daunting collection of around 60 academic papers, recommended by my committee members who specialize in diverse fields like Computational Methods in Social Neuroscience, Memory, Social Interaction and Face Perception. Initially overwhelmed by the volume and variety of the papers, I noticed overlaps in topics and references among them. To tackle this challenge, I leveraged Python, guided by Brandon Rose's tutorial on document clustering. I converted the papers from PDF to text, then used Python libraries such as NLTK and Scikit-learn for text processing and clustering. By implementing KMeans clustering and visualizing the results through dendrograms and clustermaps, I successfully categorized the papers into coherent groups. This not only streamlined my study process but also gave me insights into the interconnections between the diverse topics presented by my committee members.

Jupyter Notebook