The Loop Behind Anxiety, Avoidance, and Compulsions

Follow the Relief

Here is a scene that plays out thousands of times a day, in thousands of lives, and almost nobody recognises it for what it is.

A woman gets an invitation to a work function. The moment she reads it, something tightens in her chest. She imagines walking in alone. She imagines the small talk, the long pauses, the moment when she runs out of things to say and everyone can see she does not belong. The feeling builds. She replies “Sorry, can’t make it” — and within seconds, the tightness dissolves. Relief washes through her like a cool wave. She feels better immediately.

But here is the part she does not see: three weeks later, the next invitation arrives, and the tightness is worse. Not just a little worse. Noticeably worse. The imagined scene is more vivid, the predicted humiliation more certain, the escape more urgent. She declines again. Relief again. And again the next time it is worse.

She is not weak. She is not broken. She is caught in a loop — and the loop is operating exactly as designed. Every time she avoids, the relief teaches her brain one lesson: avoidance works. And every time she avoids, she loses an opportunity to learn a competing lesson: I can handle this. The loop tightens. The world shrinks. And the thing that feels like the solution — the decline, the escape, the quick relief — is the very thing keeping the problem alive.

Follow the relief. Wherever relief is the reward, you will find a loop.

This post is about feedback loops — the invisible engines that drive anxiety, avoidance, compulsions, reassurance-seeking, and a dozen other patterns that people spend years trying to stop through willpower alone. Once you can see the loop, you stop fighting the moment and start redesigning the system.

Two Kinds of Loop

A feedback loop is when what you do today changes what happens tomorrow. That is the entire definition. Your current behaviour alters the conditions you will face next, which alters your next behaviour, which alters the conditions again. Round and round.

There are two types, and the distinction matters enormously for understanding psychological patterns.

Reinforcing Loops: The Spiral

A reinforcing loop amplifies whatever is already happening. More of A produces more of B, which produces more of A. If the direction is upward, you get growth — confidence builds on success builds on confidence. If the direction is downward, you get a vicious cycle — avoidance breeds fear breeds avoidance.

Reinforcing loops do not self-correct. Left alone, they accelerate. They are the reason small problems become large problems, and the reason small improvements can become large improvements. The direction of the spiral depends entirely on what behaviour is being reinforced and what the reward is.

Balancing Loops: The Thermostat

A balancing loop resists change. It has a set point — a target — and it pushes the system back towards that target whenever it drifts. Your body temperature is a balancing loop: too hot and you sweat, too cold and you shiver. The goal is stability, not growth.

In psychology, balancing loops show up as the frustrating experience of making progress and then snapping back. You push yourself to socialise, start feeling better, then mysteriously stop going out again. The balancing loop has a set point for how much discomfort you are “supposed” to tolerate, and it pulls you back to that baseline. This is not self-sabotage. It is a system doing exactly what it was designed to do — maintain equilibrium. The problem is that the set point was calibrated in a context that no longer applies.

The Mechanism: Relief Reinforcement

The engine that powers most anxiety-related loops is a process I call relief reinforcement. It works like this:

1. Something triggers a threat signal. It could be a situation, a thought, an image, a bodily sensation, or even nothing identifiable at all — just a sudden spike of “something is wrong.”
2. Anxiety rises. Your nervous system shifts into alert mode. The discomfort is immediate and visceral.
3. You perform a behaviour that rapidly reduces the anxiety. You avoid. You check. You seek reassurance. You perform a ritual. You leave. You Google. You confess.
4. The anxiety drops. Fast. This drop — this contrast between the high of the distress and the low of the relief — is the reward signal. Your brain registers it as: that behaviour worked.
5. Next time the trigger appears, you are more likely to perform the same behaviour. Not because you consciously decided to, but because your learning system has filed it as the correct response.
6. And because you performed the behaviour instead of staying with the discomfort, you never collected the data that would have shown you the threat was manageable. The assumption that drove the anxiety — “this is dangerous, I cannot handle it” — remains untested and therefore unchallenged.

This is not a character flaw. It is not a lack of courage. It is operant conditioning — the most well-established learning principle in behavioural science. Any behaviour that is followed by a rapid reduction in distress will be strengthened. Your brain does not care whether the behaviour is helpful in the long run. It cares that it worked right now.

Short-term relief, long-term cost. That is the signature of every reinforcing loop running in the wrong direction.

The speed of the relief matters. A behaviour that reduces anxiety in five seconds is reinforced far more powerfully than one that reduces it in five hours. This is why avoidance and compulsions are so sticky — the payoff is almost instantaneous. And it is why the slow, steady work of building tolerance feels so unrewarding by comparison. You are competing against a system that is optimised for speed, not wisdom.

The Avoidance Loop

Let me lay out the most common anxiety loop in its full six-step structure, so you can see every moving part.

Notice the cruel precision of it. Every component serves the loop. The trigger feeds the prediction. The prediction feeds the anxiety. The anxiety feeds the avoidance. The avoidance feeds the relief. And the relief feeds the next trigger — by ensuring the fear never gets tested, never gets disconfirmed, never gets weakened by contact with reality.

The person inside this loop is not choosing to be afraid. They are caught in a system that rewards the exact behaviour that keeps them stuck.

The Reassurance Loop

Avoidance is the most visible loop, but it is not the only one. Reassurance-seeking runs on the same engine with a slightly different fuel.

Reassurance is a payday loan. You get temporary relief at an interest rate that guarantees you will need to borrow again tomorrow — and for a larger amount.

This is why reassurance-seeking is so difficult to address in relationships. The person asking is not being unreasonable — their distress is genuine, and the request does work in the moment. But the moment is all it buys. The loop ensures that the relief is temporary and the tolerance is permanently lowered. Every reassurance given is a deposit into a system that demands more reassurance.

The Mistake Everyone Makes

Here is where most people go wrong, and where most well-meaning advice goes wrong too: they treat the loop as a moral problem.

“I just need more willpower.” “I need to push through the fear.” “I should be stronger than this.” “Why can’t I just stop?”

This framing is not just unhelpful — it is structurally incorrect. It locates the problem in the person rather than in the system the person is caught in. And because the framing is wrong, the solutions it generates are wrong too. More willpower. More self-criticism. More brute-force attempts to override the loop through sheer determination.

Willpower works against the loop the way a bucket works against a rising tide. You can bail water for a while, but if you have not addressed the mechanism that is producing the water, you will exhaust yourself and the water will keep rising. The behavioural science term for this is “ego depletion” — willpower is a finite, depletable resource, and using it as your primary tool against an automated reinforcement loop is a strategy with an expiry date.

Think of it this way: if you walked into a room and the thermostat was set to 35 degrees, you would not stand in the middle of the room fanning yourself and calling it a character-building exercise. You would walk over and change the thermostat. The loops running your anxiety are thermostats. The question is not “how do I endure more heat?” The question is “where is the dial, and how do I turn it?”

Three Levers for Changing a Loop

Once you can see a loop, you have three points of intervention. You do not need to use all three at once. In fact, you should not. Pick one, apply it consistently, and let the system adjust before adding complexity. Consistency beats ambition. Always.

Lever 1: Reduce the Reward (Add Delay)

The power of relief reinforcement depends on speed. The faster the relief follows the behaviour, the stronger the reinforcement. So one lever is to slow down the payoff.

This does not mean blocking the behaviour entirely. It means inserting a gap between the urge and the action. When you feel the pull to check, to avoid, to seek reassurance — you do not say “I will never do this again.” You say “I will wait five minutes.”

Five minutes does not sound like much. But five minutes of sitting with the discomfort instead of immediately resolving it does two things: it weakens the reinforcement signal (because the relief is no longer instant), and it gives your nervous system a small experience of tolerating the uncertainty without catastrophe. That small experience is data. Over time, the data accumulates.

Start with whatever delay you can sustain without white-knuckling it. Two minutes. Sixty seconds. The number does not matter. What matters is that the gap exists and that you build it gradually. This is the principle behind find-a-five — keeping the discomfort at a level you can tolerate (around 5 out of 10) rather than flooding yourself at a 9.

Lever 2: Build a Competing Loop (Micro-Approach)

Avoidance loops are powerful because they are the only loop running. There is no competing circuit. The brain has one option filed under “what to do when anxious” and it is avoidance. So you need to build a second option — a competing loop that runs in the opposite direction.

This is what exposure therapy does, but the word “exposure” makes it sound more dramatic than it needs to be. A micro-approach is the smallest possible step towards the feared thing — so small that your anxiety stays manageable and the experience registers as “I did it and I was fine” rather than “I survived but barely.”

If the avoidance loop is about parties, the micro-approach is not going to a party. It is standing outside a busy cafe for three minutes. If the avoidance loop is about phone calls, the micro-approach is calling a business where the interaction is scripted and brief — checking opening hours, confirming an appointment. If the reassurance loop is about your partner’s mood, the micro-approach is noticing the urge to ask “Are we okay?” and instead saying “I noticed you seem quiet — how was your day?”

Each micro-approach that goes reasonably well is a deposit into the competing loop. Over time, the competing loop gains strength. It does not erase the avoidance loop — that neural pathway still exists — but it gives the brain a second option. And when two loops compete, the one that gets more repetitions wins.

Lever 3: Make the Loop Visible (Seven-Day Tracking)

The most underrated intervention is simply seeing the loop. Most loops operate below conscious awareness. You do not notice the trigger-prediction-anxiety-avoidance-relief sequence because it happens fast and it feels like a single event (“I just didn’t feel like going”). But when you slow it down and write it out, the structure becomes visible — and once it is visible, it loses some of its automatic power.

For seven days, carry a small notebook or use a notes app on your phone. Every time you notice an avoidance, a checking behaviour, a reassurance request, or a compulsion, write down three things:

1. What triggered it? (Situation, thought, sensation, or feeling.)
2. What did I do? (The specific behaviour.)
3. What was the payoff? (Usually relief, certainty, or a sense of control.)

You are not trying to change anything yet. You are collecting data. Most people are genuinely surprised by what they find — loops they had no idea were running, frequencies far higher than they estimated, payoffs they had never consciously identified. The data itself is the intervention. Once you can see the loop on paper, your relationship to it shifts from “this is just who I am” to “this is a pattern, and patterns can be changed.”

The Loop Map

Why the Discomfort Is Necessary

I want to be straightforward about this, because it is the part most people do not want to hear.

You cannot redesign a relief-based loop without temporarily losing the relief. That is not a side effect of the process. It is the process. The loop is maintained by the reward. If you change the loop, you change the reward. And changing the reward feels, for a while, like something is wrong.

This is where most people quit. They start the delay. They attempt the micro-approach. The discomfort rises — as predicted, as expected, as necessary — and they interpret the discomfort as failure. “It is not working. I feel worse. I should go back to what I was doing before.”

But the discomfort is not a signal that you are going backwards. It is the transition cost of moving from one loop to another. Every system resists redesign. Your nervous system is no exception. The question is not whether the discomfort will come — it will. The question is whether you are prepared for it, whether you have a plan for sitting with it, and whether you understand that it is temporary.

From Loops to Delays

If you have followed the logic this far, you may have noticed something: the three levers all share a common element. Delay the reward. Approach gradually. Track over time. Each one introduces time into a system that was operating on speed.

This is not a coincidence. Time is the most underestimated variable in every psychological system. The speed at which a consequence follows a behaviour determines how powerfully that behaviour is reinforced. The delay between an intervention and its effect determines whether people stick with it or abandon it. The lag between changing a pattern and feeling the benefit determines whether the change survives.

In the next post, we will look at delays as a system concept in their own right — why the gap between action and consequence is where most psychological change succeeds or fails, and how to use that understanding to stop quitting things that are actually working.

Key Takeaways

• A feedback loop is when what you do today changes what happens tomorrow. Reinforcing loops spiral; balancing loops self-correct.
• Most anxiety patterns are reinforcing loops powered by relief reinforcement — the rapid drop in distress after avoidance, checking, or reassurance.
• The avoidance loop has six steps: trigger, prediction, anxiety, avoid/escape, relief, fear generalises. Each step feeds the next.
• The reassurance loop works the same way, but the payoff is temporary certainty — and each payoff lowers your tolerance for uncertainty.
• Willpower fights the moment. Redesigning the loop changes the system. Stop bailing water and change the thermostat.
• Three levers: add delay (slow the reward), build a competing loop (micro-approach), and make the loop visible (seven-day tracking).
• Discomfort when removing relief is not failure. It is the transition cost of changing from one loop to another. It is temporary.

You do not need to break the loop in one dramatic act of courage. You need to see it clearly, pick one lever, and apply it with the consistency of someone who understands that systems change slowly and then all at once.