Throughout our lives, we often encounter events that violate our predictions about the world, or surprises. We encounter these surprises in many facets of life, from movies to breaking news to sports. In sports, surprises may be associated with positive feelings for fans – in fact, a great deal of the attraction for committed sports fans is the surprise that comes from watching teams score. These effects are fairly understood sociologically, but the psychological and neural aspects are mostly unknown. While jumping and cheering for your team after a surprising goal, did you ever wonder what happens in your brain? A new study from the Computational Memory Lab at the Princeton Neuroscience Institute led by Kenneth Norman gives us a glimpse into how our brains process these surprises.
The idea for the project came when James Antony, a CV Starr fellow at the Princeton Neuroscience Institute and a member of the Computational Memory Lab, was at the 2014 Kavli Brain Camp (Santa Barbara), which happened to coincide with the World Cup. During the days, he listened to lectures about emotional memory, and he also streamed the soccer matches in the background. He thought of how intense fans got during scores in soccer as opposed to other sports, and he thought about how, with such sparse scoring, every goal really mattered and came as a large surprise. Putting this idea together with the workshop lectures, he wondered how those surprises were processed by the brain.
Years passed while the idea fermented in James’ mind. Over the years, sports analytics became more commonplace, and sports stories were increasingly accompanied by win probability graphs that showed the crucial moments in each game when the probability of each team winning changed – and surprises occur. In parallel, he saw the groundbreaking work from the lab of Uri Hasson, another researcher at the Princeton Neuroscience Institute who commonly measures responses to naturalistic stimuli such as movies. Then the Eureka moment came: James, a sports fan himself, wondered “why don’t we use these win probability graphs to measure surprise and simultaneously put people in an fMRI scanner to measure the brain’s response to surprising moments”?
Surprise has been studied in neuroscience in well-controlled laboratory settings since at least the 1980s, when researchers used Event Related Potentials (ERPs) to study violations of expectations and found brain signals correlated with surprises. However, owing to the availability of improved technology and advanced machine learning techniques that allow one to track brain responses to naturalistic stimuli, cognitive neuroscience nowadays is gearing towards studying behaviors under conditions that are closer to real life. The sports approach, in turn, allows one to study surprise in naturalistic setting.
In the study, James and his co-authors combined fMRI, pupillometry and behavioral analysis to probe neural mechanisms underlying surprises during sports watching. In this case, the authors chose basketball because it allowed for a lot of surprises in a small amount of time. The study’s subjects were basketball fans, and they watched the final 5 min of nine games from the 2012 men’s NCAA college basketball tournament. While they were watching, their brains were scanned and their pupil diameter was measured.
The study found that the subjects’ pupils dilated during surprises, reflecting an increased level of arousal. Pupil dilation is controlled by the noradrenergic neural signaling system, which has been reported to track unexpected changes in the environment. Additionally, when surprises were positive for each subject’s preferred team, areas associated with rewards processing such as ventral tegmental area (VTA) and nucleus accumbens were activated. This shows that surprise activated regions associated with a positive affective feeling, which also show activation for numerous other pleasurable stimuli like music and food. Altogether, these findings indicate that the reward and arousal brain systems work in concert when a large change in the expectations about the environment happens.
Another crucial question was – how does the brain track the basketball game? Does it perceive the game as one chunk or in segments?
An idea may come from a psychological theory known as Event Segmentation Theory. As humans, we experience the world continuously but when we recall it, we remember it in discrete, sequential events. The central question is how we decide when those segments occur. Imagine you wake up in the morning, head to the kitchen to prepare your favorite bagel, you cut the bagel then spread the cheese on it; then you move on to make a coffee, that should depend on a different model of the world than preparing a bagel, so one moves to another segment employing a model of “coffee making”. It turns out that we tend to segment events when we experience surprises that violate our expectations. It can be challenging to quantify surprise in naturally flowing daily activities like preparing breakfast. However, sports – with their highly quantifiable predictions offering graded measures of surprise (via changes in the win probability) – offer a lucrative opportunity to do this. In the study, the authors asked a separate set of fans to watch the same games and report what they perceive as a segment. They found that the likelihood of people segmenting parts of the game was correlated with the amount of surprise at those moments – the more surprise, the more we segment. Interestingly, brain patterns were more likely to shift after high levels of surprise, aligning with the need to shift one’s understanding of the narrative of the game. Moreover, the subject’s long-lasting memories were strongest for these highly surprising moments. In all, the results suggest that surprises help us segment the world into discrete pieces and retain the moments in our memories.
This pioneering study provides a more ecologically valid design to study memories, event segmentation, and surprise using naturalistic sports viewing. It could inspire future studies to study how surprise in other real-world domains are processed by the brain.
by Ahmed El Hady