People who replay conversations for hours afterward aren’t always overthinking — they may be processing social information their nervous system flagged as unresolved

The conversation ended three hours ago. The person who had it is now in the shower, mentally running the tape again — what they said, what the other person’s face did, that pause before the subject changed. From the outside, this looks like classic overthinking. From the inside, it feels like something more specific: an itch the mind can’t reach until it gets the situation properly catalogued.

That distinction matters more than most self-help language admits.

There’s a tendency to flatten every form of mental replay into the same diagnosis — anxiety, rumination, neuroticism — and prescribe the same fix: stop doing it. But the brain doesn’t usually replay conversations at random. It replays the ones where something didn’t resolve. A tone that didn’t match the words. A compliment that arrived a beat late. A silence that registered as a question. The nervous system flagged the exchange as incomplete, and the mind is now trying to finish the file.

Stephen Porges, the neuroscientist who developed polyvagal theory, called this background scanning neuroception — the continuous, below-conscious monitoring the body does to determine whether a given social environment is safe. Porges’s specific neurophysiological claims have been contested, and that scientific debate is live. But the observable behaviour the theory tries to describe — that humans process social signals constantly and react to them before they consciously notice — is not in dispute.

What’s interesting is what happens when neuroception flags something it can’t categorise. The interaction wasn’t clearly hostile. It wasn’t clearly warm. Something landed sideways. The brain, denied a clean read, keeps the file open.

Research on social ambiguity gives this shape. A 2026 paper in Communications Psychology looking at how people respond to risk versus ambiguity found that ambiguous social information triggers distinctly different processing patterns than information with clear stakes. When the meaning is unclear, the mind doesn’t just shrug. It works harder.

This is the part the overthinking label misses. The replay isn’t always pathology. Sometimes it’s the cognitive equivalent of a tongue finding a chipped tooth — the system is checking the edge of something it can’t yet name.

That’s the adaptive version. There’s also a version that goes wrong, and it’s worth being honest about the difference.

A study of 1,717 Turkish university students published in Scientific Reports found that rumination mediated the relationship between social anxiety and internet addiction, with loneliness acting as a serial mediator. Translation: when replay loops stop resolving and start feeding on themselves, they correlate with withdrawal, avoidance, and compulsive online behaviour. The processing function breaks down. The file never closes.

So what separates productive replay from the destructive kind? The cleanest answer comes from a 2025 University of Hong Kong study published in Nature Mental Health, which examined 900 adults across Hong Kong, Guangzhou, and Fuzhou during pandemic lockdowns. The team, led by Professor Tatia M.C. Lee, used network analysis to map the connections between loneliness, rumination, and depression. Their finding was specific: it wasn’t loneliness itself that predicted depression most strongly — it was ruminating about being lonely. Thoughts focused on one’s isolation linked tightly with feelings of loneliness, and that loop maintained the depressive network.

The mechanism reveals the dividing line. Healthy replay is investigative — it’s trying to extract information. Pathological rumination is recursive — it’s processing the same emotional content without adding new data. One is the brain working through something. The other is the brain stuck rehearsing it.

People who replay conversations productively tend to be doing one of a few things. They’re checking whether they read someone correctly. They’re testing whether their own reaction was proportionate. They’re reconstructing what was actually said versus what their amygdala translated it into in the moment. There’s something forensic about it — an effort to separate the signal from the body’s reaction to the signal.

This kind of processing often gets pathologised in cultures that prize being unbothered. The art of being relaxed about everything has become a kind of social currency, and people who think carefully about interactions can come to feel defective for doing so. But the alternative — moving through social life without ever revisiting it — has its own costs. It produces people who don’t learn from misreads, don’t notice patterns, and don’t update their model of who someone actually is.

There’s also a population that replays conversations not because they’re anxious, but because they grew up in environments where social signals carried unusually high stakes. If reading the room wrong used to result in punishment, withdrawal, or volatility, the nervous system learned that close monitoring was survival behaviour. The replay is the adult version of that monitoring, running on a system that hasn’t yet received the memo that the stakes are lower now.

This is why conventional advice to simply stop the behavior tends not to work. The behaviour isn’t a bad habit. It’s a trained protective response operating on outdated threat assessments. The fix isn’t suppression. It’s recalibration.

The body keeps a vote in this. Extended exhale breathing, rhythmic movement, humming, and co-regulation with a calm person all reduce sympathetic activation regardless of what theoretical framework you attach to them. A Forbes piece on polyvagal-informed habits for daily resilience notes that these practices have documented effects on autonomic state, even as the neuroscience underneath them gets debated. What matters for someone caught in replay is that the body’s alarm state and the mind’s analytical loop are linked. Lower the alarm and the analysis loosens its grip.

A second piece in Psychology Today on the social neurobiology of excellence makes a related point: high-performing humans are not the ones with the quietest nervous systems. They’re the ones whose systems flag information accurately and then release it once it’s been processed. The release is the skill, not the absence of signal.

That reframe is useful. The goal isn’t to become someone who never replays a conversation. The goal is to become someone whose replays terminate when they’ve extracted what they can.

A few practical markers help distinguish the two modes. Productive replay tends to generate specific observations (such as noticing when someone changed the subject during a particular topic) that can be tested or filed. Unproductive rumination tends to generate abstract self-evaluation (such as believing one always makes mistakes in conversations). Productive replay ends with a hypothesis. Unproductive rumination ends with a verdict.

Productive replay can also stop. The person notices they’ve cycled through the same ground three times, recognises there’s no new information to extract, and lets it close. Unproductive rumination doesn’t have that stop signal — partly because the loop is about the self rather than the event, and the self is always available for further criticism.

One more thing worth saying. The cultural pressure to perform serenity has made people ashamed of doing the work their minds are built to do. Social information is genuinely complicated. Tones are ambiguous. People say things they don’t mean and mean things they don’t say. A brain that takes that seriously is not malfunctioning. It’s responding to the actual texture of human interaction.

What looks like overthinking from the outside is often, on the inside, a system trying to be accurate in a domain where accuracy is hard to come by. The question isn’t whether to replay. It’s whether the replay is moving toward an answer or just rehearsing the question.

People whose minds keep returning to a conversation aren’t necessarily broken, anxious, or self-absorbed. Sometimes they’re just better readers than the culture gives them credit for being — and their nervous system is asking them to finish the page before they put the book down.