The Neuroscience of Choking Under Pressure: Why the Best-Prepared Fail When It Matters Most.

Why do skilled people fail just when it matters most? Can it be prevented, and what should they do about it? Choking under pressure means performing worse than your skill allows because the stakes feel high. It is common and it strikes the well trained. Research can now trace how it happens, has found training that reduces it, and points to the steps set out in the Protocol below.

Consider surgeons. In a survey of UK surgeons, nearly nine in ten said they had felt surgical performance anxiety, and about two in three self-reported that it had hurt how they operate.^[1] These are some of the most highly trained hands in a hospital, and most of their owners say pressure makes those hands worse. Skill alone does not protect against pressure.

Skill can even be part of the problem. In one laboratory study, experienced golfers putted worse with an unusual putter that drew their attention to the stroke.^[2] Years of practice move a skill into procedural memory, where it runs without conscious thought. Pressure can pull it back into conscious control, and there it can fall apart.

Athletes tell the same story: in a survey of competitive athletes, about three in four reported choking in their last year of playing.^[3] These are common failures among trained people, not rare events among fragile ones.

For decades, researchers disagreed about the cause. One camp said worry fills working memory, the mental space that holds what you are thinking about right now, so too little is left for the task. The other said pressure turns attention onto the moves of a skill that normally runs on autopilot. Both had experimental support.

A later set of experiments, using category-learning tasks and a practised sequence of key presses, suggested that both camps were partly right, about different tasks.^[4] Outcome pressure, the weight of what rides on the result, hurt tasks that lean on working memory, by distracting. Monitoring pressure, the strain of being watched, hurt skills that normally run without step-by-step control, by drawing attention to their steps. The type of skill and how practised the performer is at it decide which route opens.

So choking may be two problems with one symptom: worse performance. This article calls them the movement route and the thinking route. The names are shorthand for these accounts, not two mapped brain circuits. Knowing which one you face comes before any intervention. In the category-learning tasks, an intervention aimed at the wrong route did not prevent the decline.^[4]

Similar concerns show up beyond sport. A review pooling three decades of studies found that students with more test anxiety score lower on standard tests, a small-to-medium effect.^[5] A survey-based study of black and Hispanic students at selective US colleges linked stereotype threat, the pressure of being judged as a member of a group, with lower grades and changes in academic effort.^[6]

In sport, surgery and school the concern looks alike: a skilled person faces high stakes and performs below their skill. The evidence differs by setting, from experiments to surveys, so it does not show one shared brain process. The rest of this article sets out what is known about how pressure disrupts skill, and what helps.

One proposed contributor is the brain's response to stress. Its threat appraisal system, a quick check of what is at stake, decides the stakes are high. Stress increases the release of chemicals that act on the prefrontal cortex, at the front of the brain: cortisol, plus two catecholamines, norepinephrine and dopamine.^[7] A major review found that even mild stress, when it feels out of control, causes a fast and marked loss of function in this part of the brain.^[7] The loss is quick, more like a switch than a slow fade. The key area is the dorsolateral prefrontal cortex, the part of the front of the brain that holds plans in mind. A rise in stress chemicals can push it past its best level and into poor function.^[8]

Later work mapped the chemistry. The planning area works best with moderate norepinephrine and dopamine, along an inverted-U dose-response curve.^[8] Norepinephrine strengthens useful signals, while dopamine filters irrelevant ones. Too much of either can weaken function.^[7][8] The inverted-U shape itself is older than this chemistry. The old curve's data came from experiments with mice.^[9][10]

How pressure disrupts a skill also depends on the kind of skill the performer is trying to use.

For expert movements, such as a golf swing, a penalty kick or a surgical stitch, one possible failure route is explicit monitoring: watching your own movement step by step. Reinvestment theory gives one account of why this hurts.^[11] Years of practice pack a complex movement into one automatic routine.^[11] The skill runs below conscious awareness, and it has to. Checking each step deliberately disrupts the timing and coordination that make it work.^[2]

When stress hits, attention can turn inward: What are my hands doing? Where is my weight? Am I following through? That inner questioning can be the problem. In the golf experiments, pressure hurt putting once players had practised it, and training while being filmed prevented that decline.^[2] A well-practised skill can become the weak point.

A study of athletes in a sport setting found the same thing: those who choke shift their attention toward specific details of their movement, as the explicit monitoring account predicts.^[12]

Thinking skills that lean on working memory, such as mental arithmetic, strategic planning and exam questions, can fail in a different way. On this account, pressure does not turn attention onto the steps; it fills working memory with anxious worry. In a study of reasoning tests, pressure hurt people with more working memory, who normally rely on it to do better.^[13] A math study found that, under pressure, people with more working memory declined only on the most demanding problems.^[14] Earlier experiments found that pressure hurt unpractised, demanding math problems; practice that let people recall the answers removed the decline.^[27]

In these accounts, the two routes break down in opposite ways. On the movement route, conscious control steps in where it should stay out. On the thinking route, anxiety takes over the working memory the task needs.

A brain-imaging study adds a twist. In a small study of a practised two-handed computer task, researchers measured functional connectivity, the working link between the prefrontal cortex and the motor cortex.^[15] Under the highest reward, a smaller rise in this link went with a larger performance drop.^[15] The brain signals were measured before movement began. The authors read this as a possible failure of top-down control.^[15] The idea of too much monitoring and this scan result come from different studies and use different measures. The behavioural studies track where attention goes; the scan tracks how strongly two brain areas link, in a different task.

Newer work looks below the surface of the brain, at single cells: in one study, rhesus monkeys made reaching movements for rewards of different sizes while researchers recorded single cells in the motor cortex.^[16] At the largest "jackpot" reward, the signals for which way to move became harder to tell apart, and the planned movement blurred before the error was complete.^[16] The failure showed in the signal before it showed in the action.

In a human brain-scan study of reaching for rewards, activity in the cerebellum and a motion-processing area while preparing to move was linked to failures under jackpot rewards.^[31] The authors suggest a change in the brain's internal model of movement, its running forecast of how a move will unfold.^[31]

That matters because choking is no longer only a psychology idea. Brain imaging and single-cell recordings now study it alongside laboratory experiments. A review describes brain rhythms associated with attention and sporting performance, such as fronto-midline theta, a wave over the front of the head.^[18] Each method looks at one piece in one task; these studies do not join their findings into one tested chain.