Social Anxiety Rewires Brain Communication During Learning, Study Finds

Social anxiety may be far more than a feeling of dread before a presentation or a party. According to new research published in JNeurosci, people with elevated levels of social anxiety show a distinctive pattern of communication between brain regions that is highly sensitive to the surrounding learning context. This heightened neural coordination appears to be a double-edged sword: it can sharpen performance when learning demands are low, yet it can actively interfere with learning once the cognitive stakes rise. The findings, presented by Hanjie Liu and Melle van der Molen of Leiden University, offer one of the most detailed glimpses yet into how the anxious brain organizes itself during real-time learning, particularly when social scrutiny enters the picture.

The study focused on a phenomenon that has long intrigued researchers in cognitive neuroscience: the tendency of socially anxious individuals to be unusually sensitive to unpleasant outcomes. Decades of behavioral work have shown that people with social anxiety do not merely fear embarrassment in the abstract; they learn from negative feedback in ways that differ measurably from their less anxious peers. This learning bias, in turn, feeds a cycle of avoidance, in which anticipated social missteps lead people to sidestep parties, meetings, classrooms, and other settings where evaluation feels possible. What remained less clear was how the brain coordinates this biased learning, and whether the coordination itself changes depending on how demanding the learning situation is.

To probe that question, Liu and van der Molen recruited 146 participants spanning a wide range of social anxiety levels, from subclinical traits to more pronounced symptoms. Each participant completed a probabilistic learning task, a type of experiment in which correct answers cannot be known with certainty and learners must instead infer patterns from repeated feedback. Crucially, the researchers manipulated the social context of performance: participants completed the task either alone or under conditions of scrutiny, in which they felt observed and evaluated. As participants learned, the team recorded brain activity, allowing them to track not just which regions were active but how those regions communicated with one another over time.

The neural signal at the center of the analysis was frontal-midline theta, a rhythmic pattern of electrical activity oscillating in the theta frequency band along the brain’s midline frontal areas. Theta oscillations in this region have long been associated with cognitive control, the mental machinery that monitors performance, registers errors, and adjusts behavior when outcomes conflict with expectations. Rather than examining this signal in isolation, the researchers examined how frontal-midline theta networks reconfigured themselves across the task, measuring the degree to which activity in frontal regions synchronized with activity in more posterior and subcortical areas as participants processed feedback and updated their choices.

The results revealed a striking pattern. People with higher levels of social anxiety displayed stronger communication between brain regions, but this increased connectivity was not constant. It was sensitive to the learning context, flexibly ramping up or down depending on whether the task was being performed alone or under scrutiny and depending on the difficulty of the learning demands at hand. In conditions where learning was relatively easy, this heightened neural coordination appeared to support performance, helping participants extract the regularities of the task and respond appropriately to feedback. But when the task became more demanding, the same elevated communication seemed to work against them, hindering the very learning it had facilitated under gentler conditions.

This context-dependent flip is what makes the study notable within the broader literature on anxiety and cognition. Many previous accounts have described anxiety as a state of hyperarousal or attentional bias, in which threat-related information captures processing resources. The new findings suggest something more nuanced: the anxious brain is not simply overactive, but differently organized, with network dynamics that are calibrated to social context in a way that can be adaptive or maladaptive depending on circumstances. In other words, the same neural machinery that helps a socially anxious person navigate a simple learning environment may become a liability when that environment grows complex.

The authors frame this as a heightened sensitivity to the social context of learning. As Liu explained in the news release accompanying the study, these findings suggest that social anxiety may be accompanied by a heightened sensitivity to the social context when learning, a sensitivity that may be helpful when learning is relatively easy but could become disadvantageous as learning demands increase, potentially disrupting how different brain regions communicate and work together. This framing positions social anxiety not as a blanket impairment of learning ability, but as a reconfiguration of the learning system itself, one whose costs and benefits depend heavily on the demands placed upon it.

The implications extend beyond the laboratory. Probabilistic learning is a reasonable proxy for many real-world situations in which people must infer patterns from incomplete and noisy feedback, from navigating workplace politics to mastering a new skill in front of others. If socially anxious individuals experience a neural reconfiguration that helps under low demands but hinders under high ones, this could help explain a familiar clinical observation: people with social anxiety often perform adequately in routine situations but struggle disproportionately when the situation is both socially evaluative and cognitively demanding, such as giving a complex presentation, taking an oral exam, or improvising in a group discussion. The neural findings provide a mechanistic account of why scrutiny and difficulty together, rather than either alone, may be particularly costly for this population.

It is worth noting that the study examined subclinical social anxiety, meaning traits distributed across the general population rather than diagnosed clinical disorder. This approach allows researchers to study anxiety-related neural variation without the confounds of medication, comorbidity, or treatment history, but it also means the findings speak most directly to the broader population rather than to severe social anxiety disorder specifically. Still, the continuous range of anxiety levels among the 146 participants strengthens the case that the observed network dynamics vary systematically with anxiety, rather than reflecting an arbitrary cutoff between anxious and non-anxious groups. The authors declared no competing financial interests, and the work was published as part of JNeurosci’s ongoing coverage of cognitive and affective neuroscience.

For researchers, the study adds to a growing appreciation that mental health conditions cannot be fully understood by examining brain regions in isolation. The dynamic interplay between frontal control systems and the wider networks they communicate with appears to be a key variable in how anxiety shapes cognition, and one that shifts fluidly with context. For the broader public, the findings offer a reframing of a familiar experience: the socially anxious brain is not simply failing to cope, but is instead running a learning strategy tuned for a different set of demands, one that excels in easy conditions and buckles under complexity. Understanding that trade-off, and the neural reconfigurations that underlie it, may eventually inform more targeted interventions, from cognitive training approaches that reduce learning demands in evaluative settings to therapeutic strategies aimed at recalibrating how the anxious brain coordinates itself under pressure.

Subject of Research: Neural basis of socially anxious learning behavior

Article Title: Examining the neural basis of socially anxious behaviors

Article References: Examining the neural basis of socially anxious behaviors. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: social anxiety, JNeurosci, frontal-midline theta, probabilistic learning, brain connectivity, cognitive control, Leiden University, neural oscillations, social scrutiny, learning context, Society for Neuroscience, anxiety