Do stress and anxiety increase cognitive load and thereby deplete working memory resources?

Why stress and anxiety can eat into working memory

Working memory is the mental workspace you use to hold and manipulate information for a few seconds—remembering a phone number while dialing, or keeping the rules of a task in mind while doing it. When you are stressed or anxious, part of that workspace gets occupied by worry, vigilance, and monitoring for threat, leaving less room for the task at hand. In a study of young adults during the war in Ukraine, elevated stress was linked to reduced accuracy and stability on an N-back task (a standard working memory test), along with slower and more variable responses—a pattern the authors describe as weakened attentional maintenance and reduced working memory efficiency under high cognitive load [3].

Anxiety appears to shift attention as well. In the same study, high anxiety was associated with shorter reaction times to neutral stimuli on a dot-probe task, which the authors interpret as increased vigilance and reduced selectivity in attention allocation [3]. In other words, anxious people may spread their attention more broadly, which can help detect threats but makes it harder to focus narrowly on a demanding task.

Threat makes faces harder to remember when the mind is already busy

Laboratory experiments that induce threat through social learning—for example, learning that a particular face or context predicts something unpleasant—show that threatening conditions increase arousal and capture neural processing resources. In one set of studies, short-term memory for faces was not affected by adverse childhood experiences or by whether the context was threatening or safe, but it was impaired by high cognitive strain [4]. Under high cognitive load, more neural processing resources were allocated toward threat-related faces, and the findings point to a decreased capacity for faces when cognitive resources are limited during potentially harmful situations [4].

This fits the resource-depletion idea: when the brain is already working hard, a threatening context can pull remaining resources toward the threat, leaving less capacity for the memory task itself. The effect was not universal—highly anxious participants seemed to benefit from threatening contexts for short-term face processing—which suggests that anxiety does not simply subtract resources but changes how they are allocated [4].

Short bursts of stress can sometimes improve performance

Not all stress is harmful. In a study of 42 pilots, participants who underwent a laboratory stress induction (the Trier Social Stress Test) showed higher anxiety scores, heart rate, and cortisol, confirming that stress was successfully induced. On a working memory task, higher memory load reduced accuracy and slowed responses for everyone, as expected. But acute stress actually increased working memory accuracy under a moderate load condition—the 1-back task—leading the authors to conclude that acute stress may not necessarily impair working memory and can even improve it under certain conditions [5].

The authors discuss this pattern in terms of emotion and motivation regulation theory, suggesting that moderate stress may mobilize attention and effort rather than drain them [5]. The practical implication is that the relationship between stress and working memory is not a straight line: mild, short-lived stress may sharpen focus, while intense or prolonged stress is more likely to overwhelm the system.

Work stress, anxiety, and the limits of training

In everyday settings, the link between stress and cognitive failure shows up clearly. A training study with students and working adults found that work demands, anxiety, and stress were related to performance in the training itself, reinforcing previous research showing that work stress and anxiety lead to cognitive failures [1]. That study did not find that these factors were linked to benefits from emotional working memory training, and its results fell short of replicating earlier transfer gains—though interference effects in an emotional Stroop task did lessen in the training group [1].

Individual differences matter too. In the threat-learning studies, highly anxious participants and those with adverse childhood experiences showed different patterns of neural processing and memory under threat, even though threat learning itself appeared intact [4]. And in the wartime stress study, moderate anxiety was associated with greater risk propensity, while high stress and anxiety were accompanied by more cautious decision-making—suggesting that the level of anxiety changes not just how much working memory is available but how people use it [3].

Sources used in this answer

Related research questions