Why closed loops are so important to your brain

How junk food and the infinite scroll take advantage of your dopamine system

On the rare occasion I realise my Mum Uniform of athleisure gear is starting to collect a few too many chubby little smudged fingerprints of paint (or sauce), I’ll often drift across to a new browser tab and visit ASOS.

Their UI has changed a bit over time, but they still have a little section at the bottom that looks something like this:

I used to love a bit of retail therapy (both online and off), but since having kids, that interest has waned. When I’m on ASOS, I’m a woman on a mission — I want to get in, find what I need, and get out.

Yet, in spite of my apathy for the task at hand, I find myself still hitting that “LOAD MORE” button, and scrolling onwards, even after completing my mission! I simply feel this strange compulsion to get to the end.

Why is getting to the end more important to my brain than the original objective?

And what if some diabolical person took away that progress bar — and made the list infinite?¹

Would my brain want to scroll forever?

This is a modern question for an ancient learning system, and it has an ominous answer.

A loop is created

Every time we experience a dopamine spike, that is our brain registering something new to learn.²

Opening the loop

When our brain has learned enough to be able to predict an upcoming process (and outcome) based on a certain stimulus — a visual cue, a thought, a feeling — a special kind of dopamine spike occurs. This spike doesn’t feel like pleasure, but instead feels like a magnetic reorientation towards something.

It’s as if our inner compass has just spun around and pointed at a new direction that we simply must follow.

This predictive spike essentially opens a loop within our brain. And there are only two ways to close it — make the prediction come true, or render it false.³

Closing the loop

Closing the loop just as the brain predicts feels satisfying — like nursing a much-looked-forward-to morning latte, or clicking that last piece of a puzzle into place.

“Mama and me finishing a camel puzzle” by A.R.

These scenarios start with a predictive spike of dopamine that orients you towards some end point — coffee, puzzle completion. And when that endpoint is reached, dopamine remains flat, because the prediction came true as expected. There is nothing new for the brain to learn.⁴

Growing expectations

Every prediction our brain makes carries with it an expectation of just how “good” the outcome will be. And how satisfying a closed loop feels depends on those expectations.

A simple low-stakes loop like brushing our teeth involves a very low expectation of value, which is why the automatic sequence of teeth brushing usually ends in mild satisfaction.

But what about larger loops like the puzzle example above?

Something triggers our desire to do it — boredom, a desire to complete something, a need to chill out. That is a predictive spike, orienting us towards the new process of completing a puzzle. A global loop is opened.

But each new step in that process is its own spike-process-outcome loop — seeing the pieces (spike), sorting the pieces (process) and clicking two together (outcome). And the two combined pieces become the predictive spike for the next process and so on.

Each new nested loop carries with it a sense of mild satisfaction when it closes, but at the same time updates expectations about the future global outcome.

More effort is expended, sunk costs increase and the value of continuing increases.⁵

Each outcome serves as a predictive spike for the next, propelling us forward, making us want to keep doing the puzzle. This is why it can feel so dissatisfying to be yanked away from a puzzle, especially the closer you are to completion. Expectations have climbed high and are suddenly not met.

By the time the puzzle lies gleaming and complete on the table, the global loop of “do puzzle” is satisfyingly closed, with nary a spike in dopamine.

Even though expectations had grown high, they were satisfyingly met.

Forcing a loop closed

When an open loop is unexpectedly closed — because the process was thwarted or the outcome unobtainable — we feel mild irritation (at best) and irate frustration (at worst).⁶

It all depends upon how high expectations had climbed.

This feeling is caused by a dopamine dip — the hole created by a violated prediction. It is the brain trying to unlearn what clearly led to an inaccurate prediction.

Ever been on the home stretch of a 1000 piece puzzle only to realise there is one darn piece missing? That’s a dip of unmatched high expectations.

Progress bars

In order to calculate the expectation of value, our brain needs to know our rate of progress. How close are we to attaining the goal?

Pursuit

Say you notice you’re suddenly hungry and feel like some scrambled eggs (spike!). You scramble the eggs (process), your expectations growing as they sizzle. They soon sit steaming and ready-to-eat in front of you (outcome).

This cycle — spike-process-outcome — beautifully describes the case of reward pursuit or goal attainment.

But guess what else follows the spike-process-outcome cycle?

Consumption… Eating the scrambled eggs.

Consumption

Just like pursuit, consumption involves a large spike at onset that opens the global loop of expectation, followed by many nested spike-process-outcome loops.⁷

You see and smell the scrambled eggs (spike!). You eat a bite (process). You taste the bite (outcome). The taste lingers in your mouth and the hunger lingers in your belly (spike!), so you take another bite (process). You taste that bite (outcome). And so on. (If the bite is unexpectedly good, the outcome will cause a “reward” spike of dopamine.)

Each outcome (whether expected or unexpectedly good) serves as a predictive spike for the next, propelling you forward, making you want to keep eating those scrambled eggs.

As they disappear, bite by bite, you feel more sated, which makes each spike hit with a little less intensity. Your egg-devouring turns more into… egg-sampling.

As time goes on, the spikes eventually fade to zero.⁸ You end up (more or less) with an empty plate and a (more or less) satisfyingly “full” belly.

Natural closure

The above example may appear a little overdone (which incidentally, is how I like my scrambled eggs), but it illustrates something crucial when it comes to rewards of food.

We have two inbuilt “progress bars” that allow our brain to know when the meal ends and the final loop is closed. The first is hunger (which decreases as the meal progresses). The second is visual — the food is literally disappearing in front of our eyes.

But what about foods that are designed to prevent us from ever feeling sated?

Engineering open loops

Known as hyper-palatable foods, these little fat-and-carb bombs have been carefully engineered to be highly addictive — to be as easy to eat as possible, their taste artificially heightened and specifically designed to fade fast, leaving you wanting more.

This instantly removes the sated “progress bar” from our brain’s arsenal.

All we’re left with is the visual cue of disappearing food.

I’m fairly certain nobody has ever stared down the barrel of an empty Pringles cylinder and thought, “I feel entirely satisfied and don’t at all wish I could have another.” (I mean, it’s right there in the slogan: “Once you pop, you can’t stop!”)

But even hyper-palatable food obeys the laws of physics — they must end. This closes the last loop in a somewhat unsatisfying way (due to hunger not being sated), but closes it nonetheless.

Now… let’s turn to social media and the infamous infinite scroll.

There is no hunger that can be sated.

No food to run out.

There is no way for our poor brain to ever find closure.

An eternally open loop

In pursuit, our brain knows when we’re making progress — it’s built into the process.

The felt cost of quitting the process increases as the outcome nears.

In consumption, our brain also knows when we’re making progress — there is at least a visual cue representing percentage of reward consumed.

The felt cost of quitting the process decreases as the end nears.

Social media takes the worst of both worlds and wraps them up in a neat little bow called (ironically), the “feed.”

The “feed” is pursuit without an end goal, consumption without a disappearing reward.

Expectations are high and stay high.

As soon as we even think about our phone, a global loop is created and stays wedged open with a giant doorstopper in the shape of an infinity symbol.

Infinite scroll

Let’s break it down, so we can appreciate just how hard our brain is working when we’re passively flicking away with our thumb.

It all starts with the big cue, the one that triggers the predictive spike to kick off the whole process of “doomscrolling”: boredom, sadness, meal times, time to kill etc.

This opens the global loop.

• Then you see the first piece of content → Spike → Nested loop opened

• You consume the content → Outcome → Nested loop closed

  • But the next piece of content is visible right beneath⁹ → Spike → Loop opened

  • You consume the next piece of content → Outcome → Loop closed

    • But you see the next… → Spike → Loop opened

    • You consume the next… → Outcome → Loop closed

      • Spike

      • Outcome

        • ad infinitum

The “feed”: a never-ending series of predictive spikes

At least when eating, the spacing between nested loops — between bites — is dictated by the natural cadence of eating. When it comes to content on your feed, the space between nested loops is compressed to mere seconds.¹⁰

But at some point, you have to stop. (Need food. Toilet. Human contact of any kind.) So you force yourself to stop, and not only leave the last predictive spike hanging, but force the entire global loop closed.

Because there is no other way to close it.

This is why we feel so frustrated, empty and guilty when we eventually stop scrolling. We’re not only riding out a rather brutal dopamine dip, but we have just spent hours neither working towards something nor working our way through something.

In the brain’s world of dopamine checks and balances, it was effort that was spent on… absolutely nothing.¹¹

The feed keeps our expectations high with no goal that will ever rise to meet them, nor consumption to naturally reduce them. They stay eternally unmet.¹²

Eye-flick-thumb-flick

Any spike-process-outcome cycle repeated enough will also result in some or all of the process being automated within our subconscious.

When it comes to a nested loop within the infinite scroll, the simple motor sequence of eyes flicking across content followed by thumb flicking upwards is no doubt etched deeply within our minds.

The spike is no longer just automatically triggering an urge to view the next piece of content. It is also automatically triggering a heavily practiced motor sequence. Before we can even register the urge, our thumb has automatically flicked the next piece of content upwards to rest in the middle of our screen, right in front of our eyes.

It’s just like autoscroll, but programmed into our brain not our feed.

When there is always a “next one” automatically placed in front of our eyeballs, is it any wonder we suddenly look up and find four hours have passed?

It’s a wonder any of us break away at all.

Getting closure

Whether we’re in pursuit or consumption mode, our brain needs to know how close we are to the end.

When all the natural benchmarks our brain usually uses to estimate progress are taken away, we’re left with an unbound process with no goal.

This is why children scream when the iPad is taken away from them.

And why I always find myself compelled to get to the last item on ASOS.

When you’re trapped in the infinite scroll and find yourself just needing “one more,” know that it’s not you.

It’s your brain working as designed in a environment it wasn’t designed for.

The only way to end an open loop that was never allowed to close naturally is to close it yourself.

It’s OK to put down your phone and ride out the dip. I promise it will end (even if the feed never does).

---

How many times a day do you find yourself trapped in the infinite scroll? How do you feel when you stop? Please share in the comments. We’ve all been there!

I’m also running my own Neuro-Engineering experiment on this problem to see if I can give my brain artificial closure when doomscrolling. Stay tuned for the results! (I promise that I’m not deliberately leaving a loop open there…)

---

If you want to read more about the dopamine spike and the opportunity it gives you to rewire your brain, you might want to check this one out:

---

References for this essay, and for the wider series, are available as a collection in the Research Library, specifically:

• Dopamine, Reward Prediction, and Learning

• Tonic vs Phasic Dopamine

• Progress, Closure, and Cost of Quitting

1. They say the poor guy who invented the infinite scroll later hated himself for what it’s done to us, and has since become an advocate for ethical technology. Good on you, Mr. Raskin.

2. In this context, “spike” means reward-prediction-error phasic changes in dopamine.

3. Or exceed expectations which causes another spike. (The brain wants to learn what just led to the better-than-expected outcome.)

4. In RPE terms, no phasic burst occurs for perfectly predicted rewards, but tonic dopamine sustains ongoing behaviour (Schultz, 2016).

5. This corresponds to rising state value under temporal-difference (TD) learning. As each step is completed without termination, the expected future reward from the current state increases, making continued engagement progressively more valuable than disengagement.

6. For those who are a fan of The Big Bang Theory, the episode where Amy tries address Sheldon’s OCD by leaving things unfinished comes to mind. Lots of loops forced closed with dips in that one.

7. These spikes are admittedly smaller in magnitude than their pursuit counterparts — less “rewriting” the program and more “running” the program.

8. Satiety reduces dopaminergic motivation via gut-brain signals (e.g., leptin), shifting from ‘wanting’ to satisfaction (Berridge, 2007, Psychopharmacology).

9. Or pulled up automatically. Auto-scroll removes the last tiny vestige of agency from the loop. There is nothing left for you to do but keep your eyes open.

10. And this is where the danger lies. Even low amplitude spikes spaced closely together can subject the dopamine system to a point where receptors must desensitise (or more permanently downregulate). We’ll explore this more in a future essay.

11. Admittedly, some value can arguably be gleaned from doomscrolling, in the form of entertainment or informational value. But unless there was some goal — comment to five tweets, search for a particular video — the overall task remains goal-less, unable to be “completed” and closure never reached.

12. This compulsion to complete is a classic example of the Zeigarnik effect (discovered by psychologist Bluma Zeigarnik in 1927), where unfinished tasks create mental tension that keeps them prominent in our awareness — exactly what infinite scroll exploits by never letting you “finish.”