“But I know this! Why couldn’t I think of it at the time!!”
I’ve said versions of this in my head, and occasionally aloud. I’ve heard if after difficult meetings, during clinical events, presentations, and negotiations. You’ve probably heard versions of it too. Basically, someone competent (like you and me, naturally) looks back at an obvious piece of information, maybe a familiar process, some straightforward option perhaps, and wonders why it wasn’t available to them when they needed it.
What I’ve then heard are explanations others make, which can rapidly become the agreed position. They panicked. They don’t know the material. They weren’t concentrating. I’ve even heard that it’s a question of personal resilience. Maybe? I mean maybe those things are in play, but probably not. Some of those are moral judgments when it’s much more a cognitive challenge.
We don’t become less intelligent under pressure. What changes is the mental workspace available to us to use the intelligence we have.
Imagine a workbench
Working memory is the small-capacity system we use to hold and manipulate information in the moment. Its working memory that lets you repeat a phone number for a few seconds until you use it. It’s working memory that lets you remember the first half of this sentence while you get to the second, so you remember the sense, idea, and point. It’s working memory that lets you compare choices, update a plan when new information arrives, imagine something for a short moment, and so on.
Think about it like this.
Long-term memory is like a warehouse. Expertise and experiences stock the warehouse. Patterns and procedures are on the shelves. Facts and figures are found there. Episodes are stored there too, along with all the stuff you’ve accumulated over the years. It’s a wonderful and glorious place.
By contrast, working memory is like a workbench, where you might have new things, and you need to bring things together and do something with them. If the bench gets smaller, nothing has vanished from the warehouse. You just have less room to work.
Here’s what we know
Shields, Samza and Yonelinas (2016) meta-analysed 51 studies (2,486 people) of acute stress and executive function. In this context, executive function means the brain’s management system that includes working memory. Stress impairs working memory. The effect was small, but reliable. Cognitive flexibility took a larger hit, but it was in only six studies so let’s hold that lightly. Stress improved response inhibition, which is the ability to stop an action. A published re-analysis challenged that last finding so let’s call it a live claim.
Collectively, it isn’t as straightforward as ‘stress makes you worse at thinking’ but, more specifically, stress trades slower deliberation for fast action.
If stress simply damaged thinking, then ALL of those functions would have decreased. They don’t. The functions that degrade are the slow, more effortful, holding-things-open ones, such as keeping several things in mind, switching frames, and filtering distraction. The function that improves is fast, and terminal, that is, stopping or committing to an action. That’s the brain prepping for deliberating less, and acting more, which is a pretty good acute threat response.
So you’re not less intelligent. Yay for you! Your brain was configured for a different job than what your meeting was asking of it.
A 2023 review by Geißler and colleagues added a layer of timing. Note, this was a review and hypothesis. They argue that two different stress systems impair working memory on two different clocks. One happens within 10 minutes of the stressor, and the other around 25 minutes after it. Impairment can thus arrive fast and also outlive the visible peak of the situation. These are the adrenaline and cortisol systems, respectively.
An implication
One significant implication of this is that appearing calm or composed is not at all a cognitive assessment, but an assessment of emotional regulation. Demeanour is easier to hold than executive function. The practical outworking of this is appearing calm should carry zero weight in fitness-for-duty decisions, and we should look to time-based and structural controls instead. Rotating people by roster not appearance. Use fresh eyes on decisions. Use decision aids that don’t rely on accurately reading someone. And don’t ask someone if they’re good to continue. As we’ve talked about before, self-assessment is often terrible.
For balance, this doesn’t mean appearing calm automatically means it’s fake, or that they are impaired. Observed calm is simply not a measure of cognition, and we shouldn’t use it as such.
The machinery bit
Your prefrontal cortex is heavily involved in working memory, attention, and flexible, goal-directed control. It’s also unusually sensitive to stress chemistry. At manageable levels, noradrenaline and dopamine help prefrontal networks operate. But if we push it harder the balance changes. The Goldilocks principle is in there: not too little, not too much, but just right. Strong stress signalling weakens the circuits that keep information online and shifts control towards faster, habitual, more reactive systems.
This isn’t a malfunction. Under immediate threat, rapid orientation and well-learned action can be exactly what you need. The problem is that modern, high-pressure work usually requires the opposite. We want you to hold multiple variables, ignore distraction, forecast consequences, inhibit first responses, and integrate information from several sources at once.
We put people under pressure, then ask the most pressure sensitive parts to do the heaviest lifting.
The risk of effort
People can often maintain performance, at least for a while, by spending more effort. The trick here is that output can look fine while they’re working with thinner and thinner margins. They take a little longer, check more often, get more irritable with interruption, and shut out peripheral information. They appear to be performing but inside, their capacity is diminishing or gone. Consequently, the next interruption or problem produces a drop that to others just looks abrupt.
An issue I have here, and which I mentioned in the reverse in the third paragraph, is that we shouldn’t necessarily describe good performance under unreasonable pressure as an individual trait, because we can then admire the person who keeps absorbing load rather than examining what the organisation is doing, and what we’re making their cognition carry.
Meetings, for example
Take a challenging and difficult executive meeting. Imagine you’re listening to a complex proposal, anticipating how the CE will react, remembering some key numbers from last month, watching the clock, deciding whether to challenge a colleague while trying to formulate a response that won’t sound overly hostile. Then someone asks you a direct question which is, itself, easy. The cognitive environment is not.
If social evaluative pressure rises, part of your system starts managing threat. How am I coming across? Am I going to look stupid? Am I going to look argumentative? All of these questions reduce the space on your workbench. The same logic applies in emergency coordination, clinical work, control rooms, and negotiations. For each, while the content differs, the bottleneck remains the same.
Focus harder isn’t the issue
Once the mechanism is clear, what we consider common advice can start to look odd. Focus. Stay across it. Keep everything in mind. Those are demands on an already under-strain system.
Instead…
The better option is to reduce what has to be held internally. Write down critical information. Externalise decision criteria. Put changing facts somewhere visible. Have someone explicitly track assumptions and unresolved questions. Separate information gathering form the final decision where possible. Reduce avoidable interruptions and, when stakes rise, shorten communication and make priorities explicit.
To be clear, this isn’t dumbing down the work. Pilots use instruments, clinicians use cognitive aids, and emergency managers use status boards or sit reps. No matter how smart you are, you still have a finite working memory.
Leaders can make the bench smaller too
It’s not just the task that creates pressure. It can be created by the room. A senior person who interrupts, demands instant answers, punishes uncertainty, or makes disagreement socially expensive is adding cognitive pressure to everyone present, and may then mistake any resulting hesitancy and narrowed thinking for lack of capability.
This is where psychological safety can help reduce the load. If part of my cognition is occupied with calculating social or personal risk, and the interpersonal cost of saying what I see, that capacity is unavailable for the technical problem. High standards and psychological safety aren’t opposites; they’re supposed to go together. The more demanding the thinking, the less sense it makes to spend limited capacity on avoiding interpersonal threat.
Design the work for the brain you have
The conclusion isn’t that people are fragile under stress, and I don’t want to create the idea that we can’t, or shouldn’t, handle reasonable pressure. Sure, some can manage more than others, but it’s mostly the opposite of fragility. Human cognition is remarkably capable and people compensate impressively. But the compensation has a cost, and we can keep building critical work around an imaginary idealised brain with unlimited working memory and effortless self-control.
So when competent people miss things under pressure, the first question shouldn’t be “what’s wrong with them?”. Ask what the task required them to hold in mind, what else was consuming their attention, and what could have been moved out of their head and into the system.
No matter how full or empty the warehouse, the workbench is small. We don’t want to make it smaller still.
Photo by Vitaly Gariev on Unsplash