Your Brain Never Expected to Regulate Itself Alone
In 2006, James Coan ran an experiment at the University of Virginia that should have changed the remote work debate. It didn't get enough attention then. It still hasn't.
He put married women in an fMRI scanner and told them they'd receive a mild electric shock. Three conditions: holding their husband's hand, holding a stranger's hand, or alone. The brain scans didn't just show who felt more afraid. They showed who was doing more work. Alone, the threat response — activity in the insula and the right dorsolateral prefrontal cortex — ran at full volume. Holding a stranger's hand, it dropped. Holding their husband's hand, it dropped dramatically. The brain hadn't just found reassurance. It had found something to borrow from. It offloaded part of its own regulatory work onto the social presence beside it.
That experiment became the empirical anchor for Social Baseline Theory. What the theory proposes has never been fully absorbed by the people deciding where knowledge workers should sit.
The Handholding Experiment and What It Actually Showed
The conventional read of Coan's experiment is emotional: being with someone you love makes scary things less scary. That reading is accurate but incomplete. The fMRI data showed something more specific — a reduction in neural resource expenditure, not just in subjective fear. The brain, in the presence of a trusted other, actually did less internal work to process the same threat.
Coan's term for this is load sharing. It's not a metaphor. The social environment carries part of the regulatory burden, and the brain — apparently anticipating this — reduces its individual expenditure accordingly. The presence of another person changes the calculation before the conversation even starts.
The study was published in 2006 as "Lending a Hand: Social Regulation of the Neural Response to Threat," co-authored with Hillary Schaefer and Richard Davidson. It established a finding that Coan would spend the next two decades building a full theoretical framework around: the nervous system is not designed to regulate in isolation. It's designed to regulate in company.
What Social Baseline Theory Actually Claims
The theory's central proposition sounds counterintuitive if you've absorbed the conventional language of resilience and independence: solo self-regulation is not the default state. It is the deviation.
The argument is evolutionary. Human beings have spent almost their entire existence as a species in persistent social environments. We didn't evolve regulatory systems for solitude. We evolved them for cooperation. The brain's baseline assumption — the state from which it measures cost and allocates resources — is the presence of others. When others are available, the brain can distribute regulatory work across the network. When they're not, it pays for everything alone.
This means that removing social presence doesn't return someone to a neutral state. It places them in a more expensive one. The brain compensates for the absent network, but compensation costs more than borrowing would have. Every regulation task — maintaining focus, processing threat, managing emotion, making judgment calls under uncertainty — runs at higher metabolic cost in the absence of social baseline.
Coan's subsequent research extended this in a direction that matters for how we design work environments: even passive social presence carries regulatory value. Not conversation. Not collaboration. Proximity. Being near another person, without interacting at all, measurably reduces the cost of self-regulation. The nervous system, observing that others are nearby, partially extends its baseline assumption — even in the absence of active social engagement.
The Async-Only Problem Nobody Is Naming
The mainstream debate about remote work is framed around collaboration, meeting fatigue, office politics, commute time, and output measurement. These are real concerns. They're also not the deepest layer.
The deeper issue is regulatory load. A fully async remote setup — no shared physical space, no ambient social presence, minimal real-time interaction — removes the distributed regulatory infrastructure the nervous system evolved to use. Every regulation task that would have been partially shared with an ambient social environment now runs entirely internally.
This doesn't manifest as loneliness, necessarily. Many fully remote workers report not feeling lonely. What they report is a particular kind of fatigue. A depletion that doesn't scale with hours worked. A flatness that arrives mid-afternoon even on light-task days. An end-of-week exhaustion that doesn't resolve with rest the way physical tiredness does.
That pattern is consistent with what Social Baseline Theory predicts: not emotional distress from isolation, but increased metabolic cost of solo operation. The brain is running more work per unit of time than it would in a social baseline environment, and the cost accumulates invisibly in productivity logs.
The knowledge work context compounds this. Knowledge work is almost entirely self-regulatory: sustained attention, inhibiting distraction, working memory, judgment under uncertainty. These tasks require continuous executive function. Stack solo self-regulation — paying for all your own regulatory load — on top of work that is already inherently self-regulatory, and the depletion multiplies.
An office worker who speaks to no one all day, who just shares a floor with other people, is still running part of their regulation on borrowed infrastructure. A fully async remote worker is running all of it internally, all day, every weekday.
What This Changes About Hybrid Decisions
The hybrid debate settles on "how many days in the office" as the operative question, and three days has become the default answer — not because anyone measured what three days actually addresses, but because it sounds like a reasonable compromise between competing preferences.
Social Baseline Theory suggests the question is wrong. The right question isn't "how many days allows sufficient collaboration?" It's "how much ambient social presence is required before the regulatory load shifts enough to matter?"
These aren't the same question. The first optimizes for coordination. The second optimizes for nervous system operating cost. You can have a day in the office full of scheduled meetings and no meaningful ambient regulation. You can have a day in the office with no meetings where the mere presence of others in the building reduces your individual regulatory expenditure for eight hours.
For organizations making hybrid decisions, this points to a calculation that almost nobody is running. The productivity model counts output. It doesn't count who paid for the self-regulation that produced it. Teams operating in fully async environments are running a hidden deficit — one that shows up as attrition and burnout long before it shows up in sprint velocity.
The Autonomy Argument's Missing Term
Neither side of the remote work debate has absorbed what Social Baseline Theory implies. The pro-office position argues from collaboration and culture. The pro-remote position argues from autonomy, commute elimination, and trust. Both are arguing about things above the level of the nervous system.
Coan's framework suggests that genuine, well-functioning autonomy requires enough regulatory baseline to operate cleanly. Removing social presence doesn't liberate the brain from regulation. It forces the brain to fund all of its own regulation, alone. That's not the same thing as freedom. It's a different kind of constraint — one that's metabolic rather than geographic, invisible rather than legible, cumulative rather than acute.
This doesn't mean remote work is wrong. It means the full cost of async-only environments has never been honestly calculated, and the missing number is metabolic, not behavioral.
Related: if the question of what remote work does to worker identity — as distinct from what it does to the nervous system — interests you, this post on autonomy grief among AI-era workers covers the identity dimension of that same shift.
The isolation cost isn't productivity. It's who pays for the regulation.
Photo by Thirdman via Pexels — two people working in shared space, the ambient social presence that keeps the nervous system from working alone.