Originally published in February 2019 and comprehensively updated/edited in June 2026.
How the late 2010s turned chronic stress into a social condition — exploring the neuroscience, social psychology, and media architecture behind an era of collective adrenaline.
For a stretch of years in the late 2010s, millions of people spent most of their waking hours in a physiological state that closely resembled a low-grade emergency. The threat wasn’t physical. It was informational. Yet the body’s response — elevated cortisol, narrowed attention, heightened reactivity — looked remarkably similar to what it would have been if the danger had been real.
In plain terms: humans were running on adrenaline.
This wasn’t a metaphor, or not entirely. The stress response systems that evolution built into the human brain were being activated at an unprecedented scale and frequency — not by predators or physical danger, but by a continuous stream of alarming news, political shocks, and social media content specifically optimized to provoke reaction. And the effects of that activation touched nearly everything: how people thought, how they argued, how they voted, how they related to one another, and how they understood themselves.
Understanding what happened during those years — roughly 2016 through 2020 — requires looking at the intersection of neuroscience, social psychology, and the specific technological conditions that made the period unlike any before it. The story isn’t simply about politics, or social media, or any single cause. It’s about what happens to human cognition and behavior when an entire civilization is placed under sustained psychological stress.
The Biology of a Society Under Stress
What the Stress Response Actually Does
To understand the late 2010s as a psychological phenomenon, it helps to start with the basics of how the human stress response works — not as a character flaw or a sign of weakness, but as a biological system doing exactly what it was designed to do.
When the brain perceives a threat — whether real or symbolic — it activates the hypothalamic-pituitary-adrenal (HPA) axis and triggers the release of stress hormones, primarily cortisol and adrenaline (also called epinephrine). These hormones prepare the body for quick action. Heart rate increases. Blood sugar rises. The immune system shifts into a short-term defensive posture. Attention narrows onto the perceived threat. All of this is the well-known “fight-or-flight” response, and it’s extraordinarily useful when you need to react to danger fast.
The problem is that this system was built for acute threats — a predator, a fall, a rival. It wasn’t built for persistent, low-grade, information-based stress that never fully resolves. When the stress response is activated repeatedly or chronically, the downstream effects on cognition are significant. Research in psychophysiology has consistently shown that chronic stress impairs functions in the prefrontal cortex — the region most associated with rational deliberation, long-term planning, and the regulation of emotional reactions. At the same time, chronic stress tends to strengthen reactivity in the amygdala, the brain’s threat-detection center.
Put simply: sustained stress makes people faster to react emotionally and slower to think through problems carefully. It shifts the brain from a mode of deliberate reasoning toward a mode of rapid threat assessment.
Why the Threat Was Informational
For most of human history, the threats that activated the stress response were environmental — predators, weather, famine, conflict. These threats were finite. You escaped, you survived, and your stress hormones eventually cleared from your bloodstream.
By the mid-2010s, a new category of persistent threat had emerged: the informational threat. This wasn’t a predator or a storm. It was a constant stream of alarming news, outrage-inducing social media posts, political developments framed as existential crises, and an endless supply of evidence that the world was dangerous and people on the other side of whatever divide mattered to you were actively malicious.
Crucially, this kind of threat never resolves. You can’t outrun a Twitter feed. You can’t fight a news cycle. The same brain systems that evolved to handle discrete physical dangers were now being activated by stimuli that refreshed every thirty seconds and never went away.
Research from the American Psychological Association’s “Stress in America” surveys during this period documented rising rates of stress linked specifically to the news and the political climate. By 2017, a significant majority of Americans reported that the news caused them stress, and a meaningful portion reported checking the news constantly despite feeling worse for doing so. This pattern — compulsive engagement with a source of distress — is consistent with what researchers call a “negativity bias loop,” in which the brain’s evolved tendency to prioritize threatening information is exploited by media systems that have learned to supply it continuously.
A Day in the Life of the Adrenaline Era
It helps to make this concrete. Picture a fairly ordinary person in 2018 — not a political activist, not a media professional, just someone with a smartphone and a social media account. They wake up and check Twitter before getting out of bed. There’s breaking news: a political development that their feed is treating as a potential catastrophe. They open a news app. Then Reddit. The headlines are alarming. By the time they’ve made coffee, their cortisol is elevated and their attention is narrowed onto a problem they have no direct ability to influence.
At work, they check their phone between tasks. At lunch, they scroll through takes and counter-takes on the morning’s news. A colleague mentions something they saw on Facebook; it’s slightly different from what they read, and the discrepancy is unsettling. By mid-afternoon they feel a diffuse, hard-to-name anxiety — not quite fear, not quite anger, somewhere between the two. They watch cable news after dinner. They lie awake at midnight replaying things they read, arguing in their head with strangers whose positions they encountered online.
The next morning, they do it again.
This was not an unusual pattern. It was, for a substantial portion of the connected population, something close to the default. And its psychological costs accumulated quietly, below the level of conscious accounting.
The Architecture of Outrage
How the Platforms Were Built
The role of social media platforms in the psychological environment of the late 2010s is not a matter of much dispute at this point, though the precise mechanisms were debated and are still being studied. What the internal research of major platforms, investigative journalism, and independent academic work has collectively established is that recommendation algorithms were designed — or evolved through iterative optimization — to maximize engagement, and that outrage-inducing content was reliably more engaging than calm content.
This wasn’t a conspiracy. It was an emergent property of optimization systems trained on human behavior. Anger, fear, and moral indignation are among the most powerful drivers of attention and interaction. A post that makes you furious is more likely to get a reaction, a share, or a comment than one that makes you mildly thoughtful. From an algorithmic standpoint, outrage is efficient.
The result was a media ecosystem in which the most emotionally activating content rose to the top of feeds, regardless of its accuracy, context, or constructive value. Users were being served a diet of content specifically selected for its capacity to produce stress responses.
The Slot Machine in Your Pocket
But outrage alone doesn’t explain why people kept coming back. Stress is unpleasant. Anger is exhausting. If social media simply made people miserable, the rational response would be to stop using it. Most people didn’t. Many couldn’t. Understanding why requires a concept that behavioral psychologists have studied since B.F. Skinner’s mid-twentieth-century work on reinforcement: intermittent reward.
Skinner discovered that animals — and humans — become most persistently attached to behaviors that are rewarded unpredictably. A lever that delivers food every time it’s pressed is satisfying. A lever that delivers food randomly, on no predictable schedule, produces something more powerful: compulsion. The unpredictability itself becomes the engine of behavior. This is why slot machines are more addictive than vending machines, even though vending machines are more reliably rewarding.
Social media feeds operated on the same principle. Pulling down to refresh a Twitter or Instagram feed is, mechanically, nearly identical to pulling a slot machine lever. Sometimes the refresh produces something rewarding — a funny post, a piece of good news, a message from a friend, a compliment on something you posted. Most of the time, it produces nothing particularly meaningful. But the occasional reward, arriving on no predictable schedule, keeps the behavior going far longer than any consistent reward would.
The neurological basis of this involves dopamine, which functions not primarily as a “pleasure chemical” as it’s often described, but as an anticipation signal. Dopamine fires in response to the possibility of reward, not just the reward itself — and it fires most intensely when the reward is uncertain. This means the act of checking a feed, not knowing what you’ll find, generates its own neurological pull entirely independent of whether the content turns out to be pleasant or upsetting. The uncertainty is the reward.
Neuroscientist Wolfram Schultz’s foundational research on what he called dopamine prediction errors helped establish this mechanism in the 1990s. His work showed that dopamine neurons respond most strongly not to rewards themselves, but to cues that a reward might be coming — and that when a predicted reward fails to arrive, dopamine activity drops sharply, motivating another attempt. Researchers studying technology behavior in the 2010s recognized the pattern immediately and applied it directly to the refresh cycle of social media use.
Former Google design ethicist Tristan Harris was among the first insiders to publicly describe this dynamic, arguing that major platforms had — whether intentionally or by optimization — built slot machines that billions of people carried in their pockets. The compulsive checking behavior wasn’t a personal failing of the people who did it. It was the expected output of a system engineered, at the neurological level, to be difficult to put down.
The Tribal Dimension
Alongside the emotional amplification, something else was happening. Social media platforms structured social life around identities — political, cultural, ideological — in ways that aligned with deep tribal psychology.
Humans are, by nature, intensely social and group-oriented animals. Research in social psychology, going back decades to the work of Henri Tajfel on social identity theory, has established that people derive a meaningful part of their sense of self from group membership. They tend to view members of their own group more favorably and to judge the behavior of out-group members more harshly, often in ways that are neither fully conscious nor easily corrected by evidence.
What social media platforms did, in effect, was supercharge this tendency. They made group identities more visible, more explicit, and more central to daily interaction. They created environments where the performance of group loyalty — loudly condemning the out-group, enthusiastically endorsing the in-group — was socially rewarded with likes, shares, and follower growth. And they made the out-group omnipresent, a constant feed of threatening ideas and infuriating behavior to react against.
The combination of chronic physiological arousal and intensified tribal psychology was, from a behavioral standpoint, a formula for significant social conflict. And that’s roughly what happened.
The Political Accelerant
When the Information Environment Met Political Instability
The psychological conditions described above did not exist in isolation. They intersected with a period of genuine and significant political upheaval in much of the democratic world. The Brexit referendum in 2016, the election of Donald Trump in the United States later that year, the rise of populist movements across Europe — these were real political events with real political causes, rooted in economic anxiety, cultural change, and institutional failures that had been building for years.
But the psychological and informational environment of the time shaped how these events were experienced and how people responded to them. The stress response doesn’t distinguish well between a real threat and a vividly described imagined threat. For tens of millions of people across the political spectrum, the political events of 2016 and after were experienced not as policy disagreements but as genuine existential dangers. The sky, depending on which feed you were reading, was always falling.
This had consequences for political behavior. Under conditions of acute threat, people tend to prefer certainty over nuance, strong leaders over deliberative processes, and in-group solidarity over across-the-aisle cooperation. This is well-documented in the psychological literature on terror management theory, which examines how mortality salience — awareness of threat and death — shifts political and social attitudes. Researchers like Sheldon Solomon, Jeff Greenberg, and Tom Pyszczynski showed across decades of work that reminders of threat and vulnerability push people toward more rigid ideological positions and stronger identification with protective groups.
The late 2010s were, in effect, a large-scale test of what happens to democratic politics when the population is under continuous low-grade threat salience. The results were not encouraging for those who value consensus, moderation, and deliberation.
The Collapse of Shared Reality
One underappreciated consequence of the adrenaline-saturated information environment was what many observers described as a collapse of shared factual reality. By the late 2010s, it had become genuinely difficult to have certain kinds of political conversations because the two parties often seemed to be living in different factual worlds — disagreeing not just about values or priorities, but about basic facts.
Some of this resulted from disinformation spread through social media. Some was the natural consequence of high-choice media environments where people selected sources that confirmed existing beliefs. But some was a direct cognitive consequence of the stress response. Research in psychology has consistently shown that people under stress are more likely to engage in motivated reasoning — interpreting ambiguous information in ways that protect group identity rather than update toward truth. Under high stress, the brain isn’t primarily in truth-seeking mode. It’s in threat-management mode.
Social psychologist Jonathan Haidt and technologist Jaron Lanier were among the most prominent voices raising alarms about where this was heading. What made the platforms particularly corrosive, in their analysis, was not simply that people encountered more conflict — it was that the specific design of social media transformed political discourse into a performance for an audience. Public posting, like counts, share metrics, and algorithmic amplification rewarded the most extreme performances most richly. A thoughtful, qualified take on a complicated issue attracted far less engagement than a hot, certain, tribe-affirming one. Over time, this shaped not just what people said publicly but arguably what they came to believe, as social feedback loops reinforced the most emotionally intense positions.
The Body Keeping Score
The experience of living through this period wasn’t just politically exhausting — it was physically costly for many people. Research from the late 2010s documented elevated rates of anxiety, depression, and stress-related symptoms in populations with high news consumption and heavy social media use. A 2018 study published in Computers in Human Behavior found that passive social media use — scrolling without actively engaging — was associated with reduced wellbeing and increased feelings of envy and social comparison. Some clinicians began informally describing a condition they called “headline anxiety”: a persistent sense of threat and helplessness generated by continuous exposure to alarming information without any corresponding ability to respond to it.
The sleep science was particularly telling. Cortisol is deeply entangled with the circadian rhythm, and its chronic elevation disrupts sleep quality. Poor sleep in turn degrades the prefrontal cortex’s capacity for emotional regulation — creating a feedback loop in which stress erodes the very cognitive resources needed to manage stress. By the late 2010s, surveys from sleep research organizations were documenting a sustained decline in reported sleep quality, with stress and screen time identified as primary contributors.
None of this is to say that people were passive victims of their neurochemistry. Millions found ways to disengage, to limit their media consumption, to maintain perspective. But the structural conditions of the period worked actively against those efforts. The economy of attention, as scholars like Tim Wu and Shoshana Zuboff were articulating in real time, had been architected specifically to prevent disengagement — and the slot-machine dynamics described above ensured that even people who wanted to stop often didn’t.
Recognizing the Pattern in Retrospect
What the Research Has Clarified
From the vantage point of 2026, some of the psychological dynamics of the late 2010s are clearer than they were in the middle of them. A substantial body of research has accumulated that helps explain what was happening at the level of both individual cognition and collective social behavior.
One of the more robust findings to emerge from this period of inquiry involves what researchers call the “outrage cycle.” Studies in political psychology and communication science found that moral outrage — the emotional response to perceived injustice or norm violation — is highly effective at driving online engagement but poorly correlated with the kinds of behaviors that actually address the underlying problem. People who felt outraged shared content, donated to causes in the immediate moment, and performed their values publicly. But sustained, organized, constructive civic action is harder to build on a foundation of constant emotional arousal than on calmer, more deliberate motivation.
Research on collective emotions — how group-level emotional states emerge and self-reinforce through social interaction — has also helped illuminate the period. Work by social psychologist Dacher Keltner and colleagues on the contagion of emotion found that emotional states, including anger and fear, spread through social networks in ways that amplify and intensify them beyond what any individual source would generate. The social media platforms of the late 2010s were, in this sense, extraordinarily efficient transmission systems for collective emotional states, capable of propagating and intensifying anxiety and outrage across millions of connected individuals in real time.
The Generation That Grew Up In It
One dimension of the adrenaline era that has become clearer with time is the particular impact it had on younger people who grew up inside it. The cohort of adolescents who came of age in the 2010s were the first generation to experience puberty — already a period of intense social sensitivity and emotional volatility — while simultaneously embedded in the social media ecosystem described above.
Haidt and researcher Jean Twenge documented in their work a significant increase in rates of anxiety, depression, and self-reported unhappiness among adolescents, particularly girls, beginning around 2012 — precisely when smartphone adoption and social media use became widespread among that age group. The causal argument is contested; some researchers argue the correlations are not as strong as claimed, or that other factors are more important. But the broad pattern — that something changed in adolescent mental health during this period — has been replicated across multiple datasets and multiple countries.
For these young people, the adrenaline era wasn’t a historical episode they passed through as adults. It was the water they developed in. The social norms, emotional habits, and political sensibilities shaped during that period have had lasting consequences for the generation now moving into civic and economic life.
The Longer Shadow
What We Learned and What Remains
It would be tidy to conclude that the adrenaline era ended cleanly and that its lessons were absorbed. The reality is messier. The fundamental architecture of the attention economy has not been dismantled, though it has been modified under regulatory pressure and shifting public awareness. The algorithmic amplification of outrage continues, albeit with more friction and more competition from platforms that have experimented with alternative engagement models.
What has changed, to some degree, is public understanding. The mechanisms described in this essay — stress response activation by informational threats, tribal amplification through social media, motivated reasoning under conditions of threat salience — are now considerably better understood by lay audiences than they were in 2016. Mental health literacy has improved. The phrase “doom-scrolling” entered common use as a recognition that people were aware of what they were doing to themselves and doing it anyway. That awareness is itself a form of cognitive resource.
What the late 2010s demonstrated, at scale and in real time, was the degree to which the human brain’s ancient threat-management architecture could be exploited — not necessarily deliberately, but systematically — by technological systems optimizing for engagement. It showed that the prefrontal cortex, the seat of deliberative reasoning and long-term thinking, is not as robustly in charge of human behavior as the Enlightenment tradition of Western culture had optimistically assumed. Under sufficient sustained stress, it yields ground to faster, more reactive systems.
It also demonstrated — and this is perhaps the most important lesson — that the social environment is not a neutral backdrop to human psychology. It is an input. The conditions under which people think, feel, and make decisions are partly technological, partly political, partly cultural, and the design of those conditions has consequences at the level of neurochemistry, social cohesion, and political behavior.
The humans of the late 2010s were not uniquely irrational or uniquely fearful. They were ordinary people with ordinary brains, placed in an extraordinary informational environment, and they responded in ways that were entirely predictable given what we know about how human minds function under stress. Understanding that is not an excuse for any of the individual or collective behaviors of the period. It’s a precondition for doing better.
Conclusion: The Aftermath of a Sustained Alarm
The years from roughly 2016 to 2020 were a kind of extended stress test — for democracies, for institutions, for communities, and for individual nervous systems. They revealed, with uncomfortable clarity, the distance between the cognitive demands of modern civic life and the evolved capacities of the human brain. They showed what happens when the architecture of communication is optimized for emotional arousal rather than understanding, and when the compulsive pull of intermittent reward keeps people tethered to that architecture even as it costs them.
The adrenaline ran out, eventually. Most people’s bodies and minds found ways to adapt, to disengage, to lower the volume. But the period left marks — accelerated polarization, eroded institutional trust, a generation of young people whose formative years coincided with the peak of the stress cycle, and open questions about the relationship between attention technology, mass psychology, and democratic governance that remain unresolved.
The informational environment has shifted since then. Short-form video replaced Twitter discourse as the dominant medium; the specific texture of the late 2010s feels dated in some ways. But the underlying dynamics — algorithmic amplification of emotional content, tribal identity reinforcement, intermittent reward driving compulsive consumption — have not gone away. They have simply taken new forms.
Perhaps the deepest lesson of the adrenaline era is this: the human nervous system never evolved to distinguish between physical danger and informational danger. It uses the same ancient machinery for both. A headline declaring civilizational collapse and a predator at the edge of the firelight activate overlapping threat-detection systems, releasing the same hormones, narrowing attention in the same ways, biasing cognition toward the same rapid-response patterns. For most of human history, that wasn’t a problem — there weren’t many false alarms, and the real threats were usually local and finite. The late 2010s created conditions that exposed just how exploitable that ancient architecture is when the “threats” arrive in an endless, curated, algorithmically optimized stream, designed by systems that had learned exactly which stimuli hit the alarm hardest.
That is not primarily a story about technology. It’s a story about what humans are, biologically — and about how much the environments we build can shape, amplify, and sometimes overwhelm the minds that have to live inside them.
