The best year of my working life was also the worst one, and for a long time I could not hold both of those facts in the same hand.
Everything landed that year. The work was good, the clients stayed, people told me I made it look easy. I remember one afternoon, a message arriving with the kind of news you are supposed to celebrate, and the first fully formed thought in my head was a list of everything that could now go wrong. Not a flicker of worry. A list. Ranked, with the worst one at the top.
I called that diligence for about ten years.
What I had wrong was the arrow’s direction. I thought the anxiety was a side effect of caring about the work. It was not. The anxiety was being paid by the work, every day, in small instalments, and I had built a career that was very good at making the payments.
This is not another list of signs. You have probably read three of those and recognised yourself in all of them, and it did not help, because recognition is not a plan. This page is about the machinery: what keeps high-functioning anxiety running when nothing is wrong, why the standard advice fails on this particular version, and what to do instead. The second half is a protocol. Skip to it if you want.
Article Summary
In one line: high-functioning anxiety persists because the behaviors that relieve it also produce results, so the way out is subtracting those behaviors rather than adding coping habits.
Most coverage of high-functioning anxiety describes what it looks like from the outside and stops there. This article addresses what keeps it running. The mechanism is negative reinforcement: over-preparing, over-checking, people-pleasing, and answering immediately each time anxiety arises within minutes, and that fast relief teaches the brain to repeat the behavior and lower the threshold for needing it.
Clinicians call these safety behaviors, and their defining property is that they prevent disconfirmation, because a fear you always pre-empt is a fear you never test.
In the high-functioning presentation, the same behaviors also produce results, so the pattern is reinforced twice, from inside by relief and from outside by praise, money and responsibility. That reframing changes what treatment means. The work is deliberate subtraction, run as behavioral experiments with predictions written down, and it requires knowing in advance that anxiety rises before it falls.
What high-functioning anxiety is
High-functioning anxiety describes a pattern where someone meets or exceeds their responsibilities while carrying constant worry, self-criticism and physical tension underneath. It is not a diagnosis in the DSM-5-TR. Clinically it usually maps to generalised anxiety disorder or social anxiety disorder. What makes it distinctive is that the coping behaviours, over-preparing, over-working, people-pleasing, relieve the anxiety and produce results, so the pattern reinforces itself.
Why it persists
Nothing in the pattern generates pressure to change. The functioning absorbs the signal, so nobody intervenes and no crisis forces the issue. The belief that the anxiety is the source of the achievement completes the lock, and it holds because the control condition has never been run: you have never done this work without the anxiety, so you have no evidence either way.
What actually helps
Cognitive behavioural therapy and SSRIs or SNRIs are first-line for generalised anxiety disorder, with good evidence for exercise and for slow breathing on acute arousal. Reassurance, better productivity systems and the evening drink all relieve in the short term and strengthen the pattern. The daily-practice layer is organised here as six areas: Regulate, Restore, Nourish, Strengthen, Connect and Realize.
What this article covers
- The relief loop that maintains the pattern, and a twelve-item inventory of the safety behaviours that run it
- The signs sorted by what other people can see against what is actually happening
- How to tell it apart from burnout, perfectionism, masked ADHD, OCD and high-functioning depression, plus the medical mimics worth ruling out first
- A self-assessment using the GAD-7, and why people in this pattern systematically under-score on it
- The evidence graded into what works, what is promising, and what makes it worse
- A four-week subtraction protocol, the six pillars that fill the space it leaves, and scripts for eight specific moments
Key takeaways
- High-functioning anxiety is not an official diagnosis, but the anxiety underneath it usually is, and it is treatable.
- The engine is relief. Over-preparing, over-checking and over-agreeing cut anxiety within minutes, and that drop teaches your brain to repeat them.
- Because those same behaviors also produce results, the loop is reinforced twice, once by relief and once by success.
- The belief that the anxiety is what makes you good is what keeps most people out of treatment, and it does not survive testing.
- Standard advice fails here because it tells you what to add. The work is subtraction, one behavior at a time.
- Anxiety goes up before it goes down when you drop a safety behaviour. Expect that, or you will quit in week one.
- A decade of low-grade threat retrains your breathing, and breathing is the one lever available in the room without anyone seeing.
What High-Functioning Anxiety Actually Is
The short definition
High-functioning anxiety is anxiety at a level that would meet clinical criteria, carried by someone whose response to it is to perform rather than to withdraw.
That distinction explains why so many people in this pattern never think the word applies to them. The textbook picture of anxiety is avoidance: someone who cannot make the call, cannot go to the party, cannot get on the plane. If that is your reference image, and you have spent twenty years making the call early, arriving first and being the calm one on the plane, you will conclude that whatever you have, it is not anxiety.
Same fear, opposite strategy. Instead of moving away from what frightens you, you move at it, harder and earlier and more thoroughly than anyone asked for. You manage fear by pre-empting it, and from the outside, pre-empting it looks exactly like competence.
It is not a diagnosis, and here is what it usually is
You will not find high-functioning anxiety in the DSM-5-TR. No clinician can diagnose you with it. Every reputable page on this subject says so and they are right.
What almost none of them say next is what it usually is. When someone with this pattern is assessed properly, the picture is most often generalized anxiety disorder, frequently social anxiety disorder, quite often both, and sometimes a persistently anxious temperament that never reaches the diagnostic threshold and still costs a great deal to run.
GAD is not rare. National Institute of Mental Health figures put past-year prevalence at 2.7% of US adults, 3.4% of women and 1.9% of men, with 5.7% experiencing it at some point in their lives, and about a third of past-year cases involving serious impairment. Those numbers count people who were assessed. They do not count the large group who never present, because they are managing.
The absence of a label has a real cost: it lets people conclude their version does not count.
The functioning is the problem, not proof that you are fine
This took me longest. I treated the functioning as evidence. Whatever this is, I thought, it cannot be serious, because look at the output.
The functioning is not evidence that it is mild. The functioning is what allows it to continue. It absorbs the signal and keeps anyone from asking. Nobody intervenes with the person who hands things in early, and no crisis arrives to force the issue, because the whole pattern exists to prevent crises. That is not a system with a natural end point. It is one that can run for twenty years, and frequently does.
The Engine: Why It Keeps Running When Nothing Is Wrong
Most anxiety makes sense to the person having it. There is a threat, they are afraid of it, and the fear is at least legible.
High-functioning anxiety does not make sense from the inside, and that is what convinces people it is a character defect. Things are going well. You have a record of handling exactly this. The dread arrives anyway, at full strength, about a meeting you will handle competently for the four hundredth time.
It makes complete sense once you look at what is reinforcing it.
Relief is a reward, and your brain is keeping score
Anxiety rises. You do the thing. Anxiety falls.
The falling is the important part, and it is the part nobody looks at.
You have a presentation on Thursday. Tuesday night the dread arrives, so you build two extra slides you will not use and rehearse the opening four times. By eleven the dread has gone. Something in you files that as a solved problem.
An email lands with a subject line that could mean anything. Your stomach drops. You open it within nine seconds. It is nothing, and the relief is immediate and physical.
Someone asks whether you can take something on, and you hear yes leave your mouth before you’ve checked your calendar, your week, or your own preference. Their potential disappointment never materialises, because you removed it.
In each case the relief arrives within minutes. That timing is close to ideal for learning. So the behaviour gets stronger, and the threshold for needing it drops, until what was a response to a genuinely high-stakes event five years ago is now standard procedure for a routine internal call.
You are not doing this because you are weak-willed. You are doing it because it works, inside the short window your nervous system can measure.
What clinicians call this
There is a name for these behaviours, and learning it reframes most of what you currently file under work ethic. They are called safety behaviours.
Paul Salkovskis described the mechanism in 1991, and work by Salkovskis, David M. Clark, Adrian Wells and colleagues built it into the standard cognitive model of how anxiety persists. The point is not that these behaviours are bad or lazy. Some are objectively useful. The point is more inconvenient than that.
Safety behaviors prevent disconfirmation.
You believe that if you turn up under-prepared you will be exposed. You never turn up under-prepared, so you never find out. The belief is never tested, never weakens, and quietly hardens into something that feels like knowledge rather than prediction. Twenty years later you are still buying insurance against a catastrophe you have no evidence for, and the reason you have no evidence is that you have never let the experiment run.
Work by Clark and Wells on social anxiety found something less comfortable still: the safety behaviors often produce the outcome the person fears. Someone avoiding eye contact and monitoring their own performance comes across as distant, and gets read as aloof. The armour causes the wound. A systematic review by Helbig-Lang and Petermann traced the same pattern across anxiety disorders generally.
The twist that makes this version so hard to escape
Here is what is different about the high-functioning presentation, and it is the most important paragraph on this page.
Your safety behaviors also work.
Not only at cutting anxiety. At producing results. Over-preparation genuinely does produce better presentations. Reliability genuinely does get you promoted. Answering within four minutes genuinely does make people want to work with you.
So the loop is reinforced twice. Once from inside, by relief. Once from outside, by praise, money, responsibility and reputation. In no other presentation of anxiety does the world applaud the mechanism while it tightens.
That is why the usual advice bounces off. When someone tells you to stop over-preparing, they are asking you to give up the thing that has, on the visible evidence, built your career. On the information available to you, refusing is rational.
A safety behavior inventory
Most people cannot see their own, because from the inside a safety behaviour does not feel like anxiety management. It feels like being a competent adult. Here is the translation.
| The behaviour | What it feels like | What it is actually doing | What it prevents you learning |
|---|---|---|---|
| Preparing far past the point of usefulness | Being thorough | Cutting pre-event anxiety | That normal preparation is enough |
| Arriving very early, every time | Being respectful | Removing the risk of arriving flustered | That arriving on time is fine |
| Rehearsing conversations in advance | Being considered | Pre-empting the fear of being caught out | That you can think on your feet |
| Re-reading messages you already sent | Being careful | Checking for the mistake that would expose you | That an imperfect message costs nothing |
| Saying yes before you have thought | Being helpful | Avoiding someone’s disappointment | That people tolerate no |
| Over-explaining, apologising pre-emptively | Being polite | Getting ahead of criticism | That you were not about to be criticised |
| Reassurance-seeking, including the casual kind | Checking in | Borrowing certainty for a few hours | That you can hold uncertainty yourself |
| Filling every gap with something | Being productive | Preventing the arrival of your own thoughts | That the quiet is survivable |
| Doing it yourself rather than asking | Having high standards | Removing the risk of being let down | That other people are mostly competent |
| Perfecting the thing nobody looks at | Caring about quality | Discharging anxiety onto something controllable | That nobody was ever going to look |
| Reviewing the whole day at night | Learning from it | Scanning for the mistake before it can ambush you | That you can leave the day where it ended |
| The evening drink that switches it off | Winding down | Shutting the alarm off chemically | That the alarm comes down on its own |
If you recognised more than four, that is normal here. When people actually count, over a week, most find between eight and twenty instances a day.
What all that worry is actually for
Worry does not feel like it has a job. It feels involuntary, like weather. But Borkovec’s cognitive avoidance model makes a case that has held up: verbal worry, the wordy, planning, rehearsing kind, blunts the physical fear response. Thinking about the disaster in sentences is less viscerally frightening than letting yourself feel it in the body. The rumination that seems to be torturing you is, at another level, protecting you from something you find worse.
Newman and Llera’s contrast avoidance model supplies the other half. Staying keyed up all day means nothing can drop on you. If you are already braced, bad news cannot produce that sickening lurch from calm to alarm, because you were never calm. Newman’s team tracked this in daily life, and the pattern held: chronic worriers are not trying to feel good; they are trying to avoid the shift from good to bad.
This is why you cannot simply decide to stop worrying, and why relaxation sometimes makes you feel worse. Coming down is not neutral for you. Coming down means you have further to fall.
Pulled outward
Look at the inventory again and notice where the attention goes. Outward, at other people’s faces, reading them for approval. Outward, at the next task. Outward, at how you are being perceived, which is a full-time surveillance job.
That is not accidental. Inward is where the alarm is, and the architecture exists to keep you from arriving there.
A dysregulated person is pulled outward. A regulated person can turn inward. That one sentence explains the busyness, the way rest feels vaguely dangerous rather than pleasant, why so many people in this pattern find seated meditation unbearable for a month and conclude they are bad at it, and why the first genuinely quiet day of a holiday is so often the worst day of the year.
The Signs, Sorted by What Other People Can See
Every other article gives you seven or eight signs and you already know them. Here is a more useful cut: the gap itself, in two columns, because the gap is the condition.
| What people see | What is actually happening | What it costs |
|---|---|---|
| Always prepared | Rehearsed it four times, because unprepared feels like exposure | Every evening before anything that matters |
| Never says no | Agreed before the sentence ended, resented it by lunchtime | A workload nobody knows is voluntary |
| Calm under pressure | Ran the disaster in full the night before, so the real thing is a repeat | Sleep |
| High standards | Cannot tell the difference between an error and a verdict on you | Every task takes far longer than it needs to |
| Very responsive | Cannot leave a message unanswered, because the gap fills with speculation | No off state, ever |
| Modest about praise | Did not hear it, heard only the clause that sounded like a caveat | Nothing good lands |
| Energetic, always busy | The gaps are the frightening part | No recovery, ever |
| Dependable, remembers everything | Holding all of it, because putting any down feels like risk | A cognitive load that never sets down |
The physical layer
Jaw. Shoulders. The band across the upper back. Headaches in the late afternoon. A stomach you have described as sensitive for so long that it is now a fact about you. Forty minutes to fall asleep on a good night. Waking at 3am with the day already running.
Most people in this pattern have quietly re-labelled all of this. It became posture, or getting older, or a sensitive gut, or just how I sleep. It stopped being a symptom and became the weather. That re-labelling is the same mechanism as everything else here: the discomfort gets absorbed rather than reported.
Eight signs the other articles do not list
- You cannot start the day until you have checked whether anything bad arrived overnight. Before coffee, sometimes before you are upright.
- Praise makes you uncomfortable and you change the subject within four seconds. Watch yourself do it. The speed is remarkable.
- You get ill on the first day of a holiday, not during the deadline.
- You apologise before you have done anything. Sorry as a conversational opening, sorry for asking a reasonable question, sorry for taking up an appointment you booked.
- You have a rehearsed answer ready for a question nobody has ever asked you, usually about why you made a decision nobody has queried.
- You would rather do it yourself than ask, even where asking would obviously be faster and the other person would obviously not mind.
- Rest requires justification. You can sit down once you have earned it, and the earning threshold keeps moving.
- You are calm in a real crisis and undone by an ambiguous two-word reply. Emergencies are strangely fine. It is the “can we talk” with no context that takes the evening.
If you are reading this about someone else
What helps: noticing effort rather than only outcomes, being specific once rather than reassuring repeatedly, and not filling their diary just because they never say no, since their yes is not reliable information about their capacity.
What makes it worse: reassurance on demand, which relieves them for two hours and strengthens the loop, and any version of “you worry too much”, which adds shame on top and gives them a second thing to manage in front of you.
Is It Anxiety, or Is It Something Else?
Several patterns produce a similar surface. The distinctions matter, because they point to different treatment.
| Pattern | The overlap | The distinguishing feature |
|---|---|---|
| Generalised anxiety disorder | Nearly total | GAD is the clinical name for what most high-functioning anxiety actually is. The difference is the label, not the condition |
| Social anxiety disorder | Fear of judgment, over-preparation, replaying conversations | The fear is specifically of being evaluated by people, and it eases in solitary work |
| Burnout | Exhaustion, dread, broken sleep, cynicism | Burnout is depleted and disengaged. Anxiety is still driven. Many people have both, in that order |
| Perfectionism | Standards, self-criticism, checking | A trait dimension, not a condition. Which kind matters, see below |
| Workaholism | Long hours, guilt at rest | Work is the object rather than the relief valve. In the anxiety version, finishing brings a brief quiet |
| ADHD, masked | Overwhelm, procrastination then panic, heavy preparation | The preparation is scaffolding for attention rather than for fear. Present since childhood, across every context |
| OCD | Checking, reassurance-seeking, intrusive doubt | Checking is tied to specific intrusive thoughts, feels ritualised, and often feels senseless to the person doing it |
| Autistic masking | Exhaustion from performance, scripts, collapse after social contact | The performance is of social convention rather than competence, and it is lifelong |
| High-functioning depression | Functioning while suffering, guilt, fatigue | The dominant tone is flat and heavy rather than keyed up. Often follows the anxiety pattern rather than competing with it |
On perfectionism, one distinction is worth having. A 2024 meta-analysis by Callaghan and colleagues, covering 416 studies and 113,118 participants, separated perfectionistic concerns, the fear of falling short and the harsh self-judgment, from perfectionistic strivings, high personal standards. Concerns showed moderate associations with depression, anxiety and OCD symptoms across clinical and non-clinical groups. Strivings showed significant but smaller ones.
That is an encouraging finding for anyone in this pattern. The high standards are not the problem. The terror of not meeting them is.
Related: the impostor experience. A systematic review by Bravata and colleagues covering 62 studies and 14,161 participants found reported prevalence ranging from 9% to 82% depending entirely on the instrument used, which tells you the measurement is unsettled. What was consistent was the company it keeps: anxiety, depression and burnout.
The medical mimics nobody mentions
Before concluding this is psychological, rule out the things that produce identical symptoms and are simpler to fix. An overactive thyroid. Your caffeine load, including the amount that used to be fine. Alcohol rebound in the early hours. Sleep apnoea. Perimenopause. Some asthma medications, decongestants and stimulants. Anaemia. Blood sugar swinging because six hours went past and you did not notice.
A GP appointment and basic bloods costs you one morning. Not one of the top-ranking articles on this subject suggests it, which I find genuinely strange.
Why the overlap does not change what you do next
The label matters for two things: which medication, and which specific therapy. Both are conversations with a professional. Almost everything in the second half of this page helps across most of these patterns.
So start, and book the appointment, rather than waiting until you have privately worked out which box you belong in. That waiting is worth examining on its own. Needing the exact right category before acting is often the pattern doing what the pattern does.
A Straight Look at Where You Are
The GAD-7
The GAD-7, developed by Spitzer and colleagues in 2006, is the standard brief screener. It is free, takes two minutes, and is used in primary care worldwide.
Over the last two weeks, how often have you been bothered by the following? Not at all = 0, Several days = 1, More than half the days = 2, Nearly every day = 3.
- Feeling nervous, anxious or on edge
- Not being able to stop or control worrying
- Worrying too much about different things
- Trouble relaxing
- Being so restless that it is hard to sit still
- Becoming easily annoyed or irritable
- Feeling afraid as if something awful might happen
Scoring: 5 or above suggests mild anxiety, 10 or above moderate, 15 or above severe. Ten is the usual threshold at which clinicians consider further assessment. This is a screening tool. Only a clinician can diagnose.
Why people like us score lower than we should
Look at items 2 and 4 again.
If you have spent twenty years building a life engineered to absorb the interference, you will answer honestly and score low. You can stop worrying, briefly, by doing three hours of work. You can relax, on Saturday, once the list is clear. The instrument asks about the experience and you are answering about the outcome, because you have made the outcome your responsibility.
Answer it again on the inside. Not on what you managed to produce despite it, but on what it took.
Five questions the GAD-7 does not ask
- If you did half the preparation, what do you actually believe would happen? Be specific. Not “it would go badly”. What, exactly?
- When did you last have an hour with nothing in it and no discomfort?
- How many hours last week went into work nobody asked for and nobody will notice?
- Who in your life knows what the inside is like? Not who is close to you. Who knows.
- If the anxiety stopped tomorrow, what are you afraid you would lose?
Question five is the one that matters most, and we come back to it.
The cost audit
Five minutes, on paper. Numbers, not feelings, because feelings can be argued with and numbers cannot.
- Hours last week spent on work beyond what was needed
- Nights of broken or short sleep this month
- Conversations you avoided because of how they might go
- Things you declined because there was no room
- Physical symptoms you are carrying and have stopped mentioning
Then look at the page. Seeing it totalled does more than any amount of self-reflection, because the pattern is very good at making each individual instance look reasonable.
When to stop reading and call someone
Book an appointment now, not after the next deadline, if any of these apply: you have had panic attacks; the anxiety has been present most days for six months or more; you are using alcohol or medication to get through evenings; your sleep has been poor for weeks; the feeling has changed from keyed up to flat, heavy or hopeless; you have had any thought of harming yourself.
If you are thinking about harming yourself, contact your local emergency number or a crisis line now. In the US, call or text 988. In the UK, call 111 or Samaritans on 116 123. Elsewhere, Find a Helpline lists services by country.
The Belief That Keeps You In It
“The anxiety is what makes me good”
Say it out loud, because most people hold it without ever having examined it.
It is not a stupid belief. It is drawn from real evidence. You are anxious, you are good at what you do, and those two facts have travelled together for as long as you can remember. Any reasonable person would draw a line between them.
What has never been run is the control condition. You have never spent a sustained period doing this work without the anxiety, so you have no data on what you would produce. None. What you have is a correlation you have lived inside for so long that it feels like physics, which is why every suggestion that you address the anxiety sounds like a suggestion that you sabotage your own career.
What is actually driving the performance
The standards are yours. The capability is yours. The fact that you care about doing the work properly is yours. None of those came from the anxiety, and none leave when it does.
What anxiety supplies is urgency and threat. Those work in short bursts. The relationship between arousal and performance is often drawn as an inverted U, and the useful part of that idea is real: some activation helps, particularly on simple, well-practised tasks. The part that gets left out is that the curve comes back down, and nobody in year eleven of chronic threat activation is sitting on top of it. You are not being sharpened any more. You have just never worked without it, so you cannot tell the edge from the erosion.
What it is costing the performance
Threat narrows attention. It biases you toward the near-term and the concrete, which helps with a deadline and is useless for a strategy. It pushes your hours toward the small perfectable task and away from the large ambiguous one, because the small task offers relief.
It makes delegation feel intolerable, and delegation is the only thing that raises your ceiling. And it steers you away from the visible, risky, career-making move in favour of flawless execution of the safe one. That effect stays invisible for years, because nothing goes wrong. It just does not go anywhere.
The anxiety does not make you excellent. It makes you reliable. Reliability has a ceiling, and most people in this pattern hit it and cannot work out why.
The second belief: that being hard on yourself keeps the standards up
Kristin Neff’s 2023 review in the Annual Review of Psychology summarises what the self-compassion literature has found: self-compassion is associated with mastery goals and with lower procrastination and self-handicapping, while personal standards stay high. Self-criticism tends to travel with more avoidance, not less. There is also a documented fear of self-compassion, a belief that easing up will cause you to slide, which predicts not practising it.
The limits are worth stating: much of that literature is correlational and the self-report measures have critics. But the direction is consistent, and it is the opposite of what the belief predicts.
Run the experiment instead of arguing about it
You cannot think your way out of a belief that has never been tested. I tried for years, and arguing with it only produces a better-argued version of the same belief. You can test it, though, on something small, this week.
What a Decade of This Does to Your Breathing
I teach breathing for a living, so treat me as an interested party and check what I say against the sources. But this section covers something all seven top-ranking articles on this subject list as a symptom and then abandon entirely.
Threat has a breathing signature
Under threat, breathing goes faster and shallower, moves up into the chest and away from the lower ribs, the exhale gets cut short, and sighs become more frequent. For a genuine emergency lasting ninety seconds, that is exactly correct.
Held for ten years, it becomes the default. A default is not something you notice. It is just how you breathe.
The loop runs in both directions
This is the part that makes it actionable. Arousal changes breathing, and breathing changes arousal.
The clearest mechanism is respiratory sinus arrhythmia. Heart rate rises slightly on the inhale and falls on the exhale, a normal and healthy oscillation. Lengthen the exhale relative to the inhale and you weight the whole cycle toward the calming branch of the nervous system. That single fact explains most of what works in this field. Zaccaro and colleagues’ 2018 systematic review covers the physiology in detail.
What the evidence supports, stated carefully
Slow breathing reliably reduces acute anxiety and physiological arousal. Balban and colleagues, 2023, found five minutes daily of cyclic sighing, a double inhale through the nose followed by a long exhale, improved mood and lowered respiratory rate more than mindfulness meditation over 28 days. Fincham and colleagues’ 2023 meta-analysis of randomised controlled trials found small to moderate effects of breathwork on self-reported stress, anxiety and depressive symptoms.
Now the limits, which matter more than the claims. These are mostly short trials with modest samples. The effects are real and they are not large. Nobody has shown that breathing alone resolves an anxiety disorder, and anyone telling you otherwise is selling something. What slow breathing does well is change your state in the next ten minutes, reliably, anywhere, for free. A smaller claim than the industry makes, and a more useful one.
Three checks, right now
- Where does the movement happen? One hand on your upper chest, one on your lower ribs at the side. Breathe normally. Which hand moves first, and which moves more?
- Is the exhale finishing? Or is the next inhale arriving while the last breath is still going out?
- Are you holding your breath right now? Reading. Concentrating. Check.
If one of those landed, that is the pattern, and it has been running since long before you noticed.
Why breathing is the right first tool for this problem
Almost everything else that helps requires visible withdrawal. Leaving the room. Taking the day. Saying the true thing out loud to someone who will then know.
Your defining constraint is that you cannot be seen to be struggling. That is not vanity, it is the core of the condition: the whole architecture exists so nobody sees the inside. Breathing is the only intervention that runs in the meeting, on the call, at your desk, in the 3am dark, with nobody able to tell.
It is a first tool, not a treatment. The next section says what the rest is.
What Actually Works, Graded Honestly
Strong evidence
Cognitive behavioural therapy. First-line psychological treatment for generalised anxiety disorder. The parts that matter most here are the behavioural ones: designing experiments, dropping safety behaviours deliberately, and building tolerance for uncertainty rather than manufacturing certainty.
SSRIs and SNRIs. First-line medication for GAD, taking several weeks to reach effect. I teach breathing and I will still say plainly: where medication is indicated, take the medication. Anyone in the wellness field who implies otherwise is putting their business model ahead of your outcome.
Exercise. Singh and colleagues, 2023, in the British Journal of Sports Medicine, reviewed 97 systematic reviews covering more than a thousand trials and found physical activity effective for reducing anxiety and depressive symptoms, with higher-intensity activity associated with larger effects.
Slow breathing, for acute arousal, as covered above.
Promising, or effective but limited
Mindfulness-based programs. A reasonable evidence base, with one caveat: sitting still with attention turned inward is frequently hard early in this pattern, for the reasons above. If your first attempts feel worse rather than better, that is expected, and starting with movement or breathing is a legitimate route in.
Worry postponement. When a worry arrives during the day, note it and defer it to a fixed 30-minute window, same time and place, usually evening. Dippel, Brosschot and Verkuil’s 2023 meta-analysis found seven trials with 999 participants and significant but small effects, around d = 0.31 for worry duration, with no long-term follow-up. Small effect, near-zero cost.
Self-compassion training, promising for the self-criticism component specifically. CBT for insomnia, where sleep onset or the early-hours wake is a main feature, with better evidence than sleep hygiene advice.
What does not work, and what makes it worse
This is the section nobody writes, and it is the most useful one here.
Reassurance. Asked for or given. It relieves for a few hours and strengthens the loop every time. The clearest tell that you are reassurance-seeking rather than gathering information is that you already know the answer and you are asking anyway.
Better systems and more organisation. The classic trap for this reader. A new productivity method feels like progress and is very often the safety behaviour in a smarter jacket. If your response to feeling anxious about your workload is to redesign your task system, notice what just happened.
“Just relax”, and forced relaxation. For some people deliberate relaxation increases anxiety rather than lowering it, and the contrast avoidance model explains why: coming down means further to fall. If this is you, start with movement, or with a breathing practice that has something to do.
The evening drink. It genuinely does switch the alarm off, which is why it is so common here. Then it fragments the second half of the night and produces a rebound in the early hours. The 3am wake you have been blaming on stress is frequently the drink, and two dry evenings will tell you.
Quitting the job. Sometimes the job really is the problem. Usually the loop travels and reinstalls itself in the new one inside a quarter. The test: does the dread attach to specific, changeable features of this role, or does it re-form around whatever is in front of you?
More willpower. It is what got you here.
How to read the evidence in this field
Small samples are the norm. Waitlist controls make effects look larger than active comparisons would. You cannot blind someone to whether they are doing breathing exercises, so expectancy effects are unavoidable. And acute effects are much better established than lasting trait change. The honest summary is “helps, modestly, reliably” rather than “transforms”, and you should be suspicious of anyone whose summary is stronger than the studies.
The Subtraction Protocol
The principle
Add nothing until you have subtracted something.
Every instinct will be to bolt a new routine onto the existing load. A morning practice, a journal, a course. Adding is what you are good at. It is also the disease presenting itself as the cure, because a new regime gives you fresh relief through fresh activity and changes nothing structural.
The work is removal. One behaviour at a time, with the discomfort planned for.
How to run a behavioural experiment
The core skill, taken directly from how CBT operates.
- Pick one safety behaviour. The smallest, not the most important.
- Write down what you predict will happen. Specifically, with a percentage. Not “it will go badly”.
- Drop it once, on something genuinely low-stakes.
- Write down what actually happened, within the hour.
- Compare the prediction with the outcome.
- Repeat until the prediction changes, usually three to five times for a small behaviour.
The writing down is not busywork. Without it, memory rewrites the outcome to fit the belief within about two days. You will remember it was fine “because it was low stakes”, or “because I got lucky”, and the belief survives intact. On paper it cannot do that.
The spike, and the curve
This is the most important thing on this page and the reason most people fail within a week.
When you drop a safety behaviour, your anxiety goes up first.
Usually for twenty to forty minutes, sometimes most of a day. It can be sharp, and it will feel like confirmation that you needed the behaviour after all.
It is not. It is what the removal of relief feels like. The behaviour was suppressing anxiety, you have stopped suppressing it, and the level you are now feeling is the level that was always there. It settles on its own, without you doing anything, and it settles a little lower each time you repeat the experiment.
If you know the spike is coming, you can sit through it. If you do not, you will read it as evidence and go back, adding one more data point to the belief that you cannot cope without the behaviour. That is the entire difference between this working and not working.
Where to start
| Behaviour to drop | The smallest possible experiment | Write this down first |
|---|---|---|
| Over-preparing | Prepare to 70% for one low-stakes meeting | What specifically goes wrong, and how likely |
| Arriving very early | Arrive five minutes before, once | What you expect to feel, and what others will notice |
| Re-reading sent messages | Send one routine email and do not reopen it | What error you expect to find |
| The instant yes | Say “let me check and come back to you tomorrow”, once | How you expect them to react |
| Pre-emptive apology | Send one message with the apology deleted | How it will be received |
| Reassurance-seeking | Sit with one uncertainty for two hours before asking | How bad it will get |
| Filling every gap | Twenty minutes with nothing scheduled | What arrives in the quiet |
| Doing it yourself | Delegate one small thing and do not check it | How badly it comes back |
| Perfecting the invisible | Ship the internal document at good enough | Who comments |
| The night-time review | Write tomorrow’s three items at 6pm and close the day | Whether you forget something |
| The evening drink | Two consecutive evenings without it | How the evening goes, and how you sleep |
| Answering immediately | Leave one non-urgent message for four hours | What they assume about you |
Four weeks
Week 1. See it. Change nothing. Log every safety behaviour as it happens, with your anxiety before and after on a 0 to 10 scale. Add the three breathing checks once a day. That is the whole week, and most people are genuinely shocked by the count.
Week 2. One experiment. One behaviour, three repetitions, predictions written down each time. Add the extended exhale as your in-the-moment tool: breathe in through the nose normally, then let the exhale run longer than the inhale, without forcing it, for six to ten breaths.
Week 3. Subtract and protect. Start a second experiment. Then protect one recovery slot in the week and defend it the way you would defend a client meeting, because the loop will negotiate for it and will have excellent reasons.
Week 4. Build. Only now do you add. This is where the pillars in the next section come in, and it is deliberate that they come last.
Expect worse before better in the first fortnight, and no dramatic moment, ever. You are primed to look for one and its absence will read as failure. The first real sign is small and specific: something that would have taken your whole evening takes twenty minutes.
What Fills the Space: The Six Pillars
Subtraction leaves a hole. If nothing goes into it the loop comes back inside a fortnight, because the loop was doing a job and the job is still vacant.
This is the structure I use, six areas, in this order, because they build on each other. Underneath all of them sits the idea from earlier: a dysregulated person is pulled outward, a regulated person can turn inward. Every pillar either lowers the outward pull or makes the inward turn survivable.
Each comes with a trap, because a high-functioning anxious person can turn any of these into another safety behaviour, and usually does.
1. Regulate: the breath
The extended exhale as your daytime tool. Nasal breathing as the default, in and out, day and night. Then, once that is established, equal-length nasal breathing gradually slowed toward five or six breaths a minute, with smoothness before slowness as the governing rule. If the breath is jerky or strained it is too slow, and a strained slow breath does the opposite of what you want.
The trap: turning it into a target, a streak and a number. The moment it becomes a metric it becomes another thing to fail at, and you have recruited the practice into the pattern.
This week: three minutes, once a day, at a fixed time. Do not measure it.
2. Restore: sleep and recovery
The 3am wake, explained: your alarm system runs a scan once the evening’s distractions are gone, and alcohol makes that scan reliable rather than occasional. Sleep-onset rumination is the same mechanism at the other end, which is why the mind starts working the instant the room goes quiet. Underneath both sits a recovery deficit built over years, because no gap has ever gone unfilled.
The trap: optimising sleep. Tracking it, scoring it, then lying awake anxious about the score. Sleep is uniquely bad at responding to effort.
This week: one 20-minute unfilled gap, in daylight, without your phone.
3. Nourish: what you are running on
Caffeine first, because it is the highest-leverage item and the least examined. The dose that used to be fine is often not fine now, particularly after a poor night, and it builds a loop of its own: bad sleep, more coffee, more arousal, worse sleep. Then alcohol as the evening off-switch and the rebound it produces. Then long gaps between meals, and the adrenaline that quietly covers them.
The trap: converting this into another rule set to be perfect about, at which point your nutrition becomes a fresh arena for self-criticism.
This week: last coffee before noon, seven days, and watch what happens to the 3am wake.
4. Strengthen: movement that discharges
The evidence here is strong and the framing for this reader is specific: movement is for discharging arousal, not for another performance metric. Which is why walking counts, and why sessions that leave you wrung out, wired and unable to settle at 10pm may be feeding the pattern rather than treating it.
The trap: turning exercise into the highest-stakes arena of all. This is the most common version among high achievers, and it is easy to spot: the training has splits, targets, a streak, and a sense of failure attached.
This week: one walk with no headphones, no route and no recording of it.
5. Connect: being known rather than being useful
The hardest pillar here and the one with the most leverage. The mask is the core of the condition. The whole architecture exists so that nobody sees the inside, which means being useful has become the way you stay in relationship with people. That works, it is exhausting, and it leaves you surrounded by people who like you and do not know you.
One distinction matters more than anything else. Reassurance-seeking feeds the loop. Being known dissolves it. The difference is the sentence. “Do you think it will be okay?” is reassurance, and the relief lasts two hours. “This is what it is actually like inside my head” is being known, and it does not wear off, because nothing was borrowed.
The trap: performing vulnerability. Being very good at talking about your struggles, in a way that is polished and costs nothing.
This week: tell one person one true sentence about the inside, with no request attached.
6. Realize: who you are without it
The last pillar and the real barrier. Question five of the self-assessment was the important one: if the anxiety stopped tomorrow, what are you afraid you would lose?
For most people the honest answer is their identity as the one who can be relied on. Which means recovery threatens something more central than comfort, and that is why people sabotage it at exactly the point it starts working. The loop’s final argument is that you are becoming someone who does not care as much.
The answer to that is not another technique. It is knowing what you actually value, separately from what you are praised for, so there is something underneath the achievement to stand on.
The trap: turning self-realisation into a project with milestones and a reading list.
This week: name three things you would still care about if nobody were watching.
The order is not arbitrary. Someone who cannot sleep cannot regulate. Someone who cannot regulate cannot be honest with anyone. Someone who has never been honest with anyone cannot ask who they are underneath all of it.
In the Moment: Eight Situations and What to Do
3am, wide awake, running the day
Your alarm system is scanning, and your brain is chemically poor at problem-solving at this hour. So do not solve anything. Long slow exhales, longer than the inhales, no counting. Feet warm. No screen. If a thought is genuinely load-bearing, write two lines on paper and put it down. The rest is the scan, and the scan does not need your participation.
Sunday evening
Anticipatory anxiety pointed at a week that has not happened. It is a rehearsal, not a forecast, and it has a poor accuracy record you can check. Three items on paper for Monday morning, then the day is closed. The dread will still be there with nothing to attach to.
The ten minutes before something that matters
Six to ten breaths with a long exhale, through the nose. It works, it is invisible, and it takes ninety seconds. Then notice the pull toward one last check of the material. That check is the safety behaviour, and skipping it is an excellent low-cost place to run the experiment.
The message that lands at 5pm
The trigger is ambiguity, not content. Notice that you have already written the worst version in your head and are reacting to your own draft. If it is not urgent, it waits four hours. Write down what you assume it says, then compare when you open it. Do this five times and the assumption loses its authority.
Someone asks you for something
The most valuable sentence here: “Let me check what I have on and come back to you tomorrow.” Twelve words that break the automatic yes and cost nothing socially. If the answer is no, two sentences with no justification. “I can’t take that on this month. I hope it goes well.” The justification is the apology in disguise, and it invites negotiation.
You made a real mistake
Separate the correction from the verdict. Fix it, say it once, plainly, stop. Then watch for the urge to over-apologise, which is the strongest one here. Over-apologising converts your error into their reassurance task, which is why a small mistake sometimes damages a relationship more than it should.
Someone praises you
Do not deflect. Do not add the caveat. Do not explain who else helped. Two seconds of silence, then thank you. Nothing else. Then notice what those two seconds felt like, because it is the most informative thing you will feel all week.
An empty afternoon
The hardest one, which is why it is last. Nothing scheduled, nothing owed. The discomfort arrives within about fifteen minutes, it is surprisingly strong, and it will present you with an extremely reasonable list of useful things to do.
Stay in it. It peaks and then it falls, with no technique at all. Sitting through it once is worth more than everything else on this list, because it is direct evidence that the quiet is survivable, and the quiet is what the pattern was built to avoid.
At Work, Where Most of This Lives
The reliability trap. You became the person who absorbs the overflow, and the reward for absorbing it was more of it. Every system routes work toward whoever does not push back. It will not correct itself, because from the organisation’s point of view nothing is broken.
Saying no without the essay. Two sentences, no reasons. “I don’t have capacity for that this month.” “I can take it if X moves, otherwise not this time.” Reasons invite negotiation, apologies invite reassurance you will then have to manage. The short version feels rude the first three times and is not.
What to tell a manager. Usually safe and useful: capacity, priorities, and what will slip if something is added, framed as information rather than confession. Often unwise: a diagnostic label, in a workplace where it will be misread. This varies by jurisdiction, employer and manager. Occupational health and employee assistance programmes are typically confidential and are a lower-risk first stop.
The promotion question. The fear is that a bigger role means more of this, and sometimes that is true. What actually changes is whether the role rewards delegation, and delegation is the safety behaviour you find hardest to give up. That is why the next level is so often where this pattern breaks: not because the work is harder, but because a job that cannot be done alone meets a person who has never let anything go.
Getting Help, and What to Say When You Look Fine
The threshold is lower than you think. You do not need to be failing to qualify. Distress is the criterion, not visible dysfunction. “I am managing, and it is costing me too much” is a complete and legitimate reason to book an appointment.
What CBT actually involves. Probably not what you are imagining. Structured and time-limited, typically twelve to twenty sessions, with a clear focus and homework between them. For anxiety, much of that homework is the behavioural experiments described above. If you have run even one, you have already started.
Medication, without the flinch. SSRIs and SNRIs are first-line for GAD. They take several weeks to work, early side effects are common and usually settle, and taking them is not a life sentence. The specific fear here is worth naming: that medication will flatten the drive your career depends on. Some people report emotional blunting and some do not, doses and drugs vary, and it is reversible. Take the question to the prescriber rather than letting it quietly make the decision for you.
What to say in the first appointment. You will walk in looking well, speak clearly and minimise, because that is what you have practised for twenty years. So write it down beforehand and read it out. Four lines: how long it has been going on; what it costs you, using the numbers from the cost audit; what you do to manage it, naming the safety behaviours specifically; and what you have stopped doing or turned down because of it. That last line usually changes the conversation, because it is the only evidence of impairment a high-functioning person can offer.
How to choose a therapist. Ask whether they work with anxiety specifically, whether they use behavioural experiments and exposure, how sessions are structured, and when you will review progress together. Reasonable answers are concrete. Warning signs: no structure, no homework, no review point, and an open-ended arrangement with no way of telling whether it is working.
What Recovery Actually Looks Like
Not calm. Not the absence of anxiety.
The gap between the outside and the inside gets smaller. You notice the urge to over-prepare and it becomes a choice rather than a reflex. Someone is disappointed in you and the day survives it. An hour opens up with nothing in it and nothing rushes in to fill it. You are still the person who cares about doing the work properly, the standards are still yours, and you have stopped paying for them with every evening.
It comes back under load. That is normal and it is not a relapse. Recognizing it in two days instead of two months is the win.
Before you close this tab: is your exhale finishing, or is the next breath arriving early?
Frequently Asked Questions
Is high-functioning anxiety a real condition?
The experience is real. The term is not a clinical diagnosis. What it usually describes is generalised anxiety disorder or social anxiety disorder in someone who copes by performing rather than avoiding, and those are real, diagnosable and treatable.
Is high-functioning anxiety in the DSM-5?
No. It appears nowhere in the DSM-5-TR and no clinician can diagnose it. That does not mean the anxiety underneath is undiagnosable, and most people who identify with the term meet criteria for something that is.
What is the difference between high-functioning anxiety and generalised anxiety disorder?
Mostly the label. GAD is the clinical category. High-functioning anxiety is a popular description of one way GAD presents, in someone whose coping response is to over-perform. If you have the pattern, GAD is the most likely formal diagnosis.
Can you have high-functioning anxiety and be genuinely happy?
Yes, in patches, and that is part of why it goes unaddressed. The problem is not that good things never happen. It is that they do not land, because attention is already on the next thing that could go wrong.
Is there a test for high-functioning anxiety?
Not specifically, since it is not a diagnosis. The GAD-7 is the standard brief screener for the anxiety underneath it and is included above. People in this pattern tend to under-score, because several items ask about interference and they have built lives that absorb interference.
What causes high-functioning anxiety?
The usual contributors to anxiety: genetics, temperament, early environment, sustained stress. What makes this version persist is separate from what started it. It persists because the coping behaviours produce fast relief and good results, so they are reinforced twice.
Does high-functioning anxiety go away on its own?
Rarely, because nothing in the pattern generates pressure to change. It typically continues until something forces the issue, often burnout, a health problem or a drop into depression. Treated, it responds well.
Is high-functioning anxiety the same as being a perfectionist?
No. Research separates high personal standards from the fear of falling short, and the standards on their own are associated with much less distress. The fear, the self-judgment and the sense that a mistake is a verdict are what track with anxiety.
Is it burnout or high-functioning anxiety?
Burnout is depleted and disengaged. Anxiety is still driven and still caring intensely. They frequently arrive in sequence, with years of anxiety producing the exhaustion that becomes burnout, so having both is common.
Can high-functioning anxiety turn into depression?
It often precedes it. Years of sustained arousal with no recovery is a recognised route into a depressive episode, and the shift from keyed up to flat and heavy is a signal to seek help promptly rather than waiting.
Will treating my anxiety make me less successful?
This is the belief that keeps most people out of treatment. The standards, the capability and the care about the work are yours and do not leave. What goes is the threat, and threat narrows thinking, blocks delegation and steers you toward safe work.
Why do I feel worse when I try to relax?
Because coming down means further to fall. Staying keyed up prevents the lurch from calm to alarm, which some people find harder than constant tension. If forced relaxation raises your anxiety, start with movement or with a breathing practice that has something to do.
Why do I crash when I finally take a holiday?
You have been running on sustained arousal that was masking how depleted you are. Remove the demand and the masking stops. Getting ill or flattened in the first days of a break is one of the clearest signals that the pattern is carrying a real physiological cost.
Why do I wake at 3am?
Your alarm system runs a scan once the day’s distractions are gone. Alcohol makes it far more reliable by fragmenting the second half of the night. If you drink in the evening, two dry nights will tell you how much is stress and how much is the drink.
Does breathwork help high-functioning anxiety?
For acute arousal, yes, reliably, with a decent evidence base for slow breathing. For the disorder as a whole it is one component, not a treatment. Its particular value here is that it works in the room, without withdrawing, which matters when your constraint is that you cannot be seen struggling.
What is the fastest way to calm down in the moment?
Lengthen the exhale. Breathe in normally through the nose, then let the out-breath run longer than the in-breath, without forcing it, for six to ten breaths. No counting needed. It takes about ninety seconds and nobody can tell you are doing it.
Do I need medication?
That is a conversation with a doctor, not a decision to make from an article. Medication is first-line for GAD and works well for many people, therapy works well for many people, and the combination is often better than either. Severity, preference and history all matter.
How long does it take to get better?
Behavioural experiments start shifting specific beliefs within weeks. Broader change usually takes months. Nothing about it is dramatic, and the first real sign is small: a reaction that used to take an evening takes twenty minutes.
How do I explain this to my partner?
Describe the gap. What they see, and what is happening underneath it. Then be specific about what helps, which is usually being noticed rather than being reassured, and about what does not, which is repeated reassurance and any version of “you worry too much”.
What should I do first?
Pick the smallest safety behaviour from the inventory. Write down exactly what you predict will happen without it. Drop it once this week on something low-stakes, and write down what actually happened. That single experiment is worth more than any amount of reading.
Where to Start
If I could get one sentence back to myself ten years earlier, it would be this: the relief is the problem.
Not the worry, not the standards, not the hours. The relief. Every time I made the anxiety go away by doing something, I taught it to come back a little sooner and a little louder, and I called that being good at my job.
So do not start with a new routine. Start with one subtraction. Pick a behaviour from the inventory, the smallest one. Write down what you believe will happen without it. Then find out this week, and write down what actually happened.
That is the whole method. Everything else here is support for that one move.
*This article is for information only and is not medical advice, diagnosis or treatment. If you are struggling, speak to a doctor or a qualified mental health professional. If you are in crisis, contact your local emergency number, call or text 988 in the US, call 111 or Samaritans on 116 123 in the UK, or find a service in your country at findahelpline.com.*
About the author: Danail Donchev: lived experience of the pattern described here, years of teaching breathing and pranayama, founder of the Pranadan Method. Scope of practice: I teach breathing and lifestyle practice. I am not a doctor, psychologist or psychiatrist, and nothing here replaces their care.