Why Stress Makes It Hard to Connect the Dots

A new brain-imaging study pinpoints the kind of learning stress pulls apart, and why it may be part of what keeps anxiety stuck.

Stress has a well-earned reputation for wrecking memory. Most of us can summon the feeling without trying: the perfectly prepared retort that disappears from our mind the second it is needed, or the material you crammed the night before a deadline that turns out, when you finally reach for it, to have barely stuck.

What is going on in our brains, when stress is getting in the way of learning? And perhaps more interesting: does stress affect our ability to fluidly associate things, connect dots and even be creative?

We all probably know the answer from our own experience. Yes. But how?

A new neuroimaging study might have some answers. First some background.

You see a friend’s new scooter, a particular shade of light blue, parked outside the university library, and without any effort at all you land on the thought that she is probably inside studying. Nobody handed you that fact. You stitched it together from two separate experiences (your friend with the scooter, and the scooter at the library) that happened to share an overlapping piece, and your brain wove them into a conclusion about a place you cannot even see. We do this constantly, so fluidly that it feels like nothing at all, and it is the difference between a mind that merely records and a mind that builds a working model of the world. The region doing most of that weaving is the hippocampus, and the process has a name: memory integration.

Neat trick!

This matters well beyond friends and scooters. Building cohesive structures out of related experiences is how a student turns scattered facts into understanding, how a witness assembles a coherent account from glimpses, and how any of us lean on the past to make sense of the present. When the weaving breaks down it tends to show up in places we care about a great deal, including psychosis and the anxiety disorders, where trouble linking related experiences is part of the clinical picture.

That leaves an obvious question, given how much of modern life is spent at least mildly stressed. We already know from decades of research that acute stress is a powerful editor of memory, reliably strengthening the laying-down of new memories while making it harder to retrieve old ones. And we know the hippocampus is studded with receptors for the very stress hormones that flood the system when we are under pressure (glucocorticoids such as cortisol). What we have understood far less well is what stress does to this more flexible, integrative work, the weaving itself. Does being stressed change our ability to braid separate experiences together and infer the things nobody ever told us directly?

That is the question a group led by Kai Schüren and Lars Schwabe in Hamburg set out to answer, and they just published their results in Science Advances. The design is clean enough to describe over coffee or tea. On day one, people learned a set of pairs (call them A and B, a face and an animal). The next day, half of them went through the Trier Social Stress Test, which is the gold-standard way researchers manufacture acute stress in a lab and is exactly as unpleasant as it sounds: an impromptu job-interview speech and some backward arithmetic, performed in front of a stone-faced panel in white coats while a camera films you. And don’t forget the bright lights shining in your face! Right afterward, everyone learned a second set of pairs that overlapped with the first (B and C, the same animal now linked to an object). Then, an hour later, came the twist nobody had warned them about: could they infer a link they had never been taught, the one between A and C?

So what happened?

Both groups learned the A-B pairs equally well, both learned the B-C pairs equally well, and both could recall those pairs on the final test with no group difference at all, so the raw material was plainly intact. What stress degraded was the inference, the capacity to put the pieces together into something larger than either piece alone (A linked to C). When the researchers used fMRI to look at participants related brain activity, they saw why: during that second learning session, the stressed brains were reactivating the old, related memory far less than the unstressed brains were, and the less a given person’s hippocampus lit up with that old memory, the worse they did at the inference later. Nothing had been erased; what stress had done was stop the brain from holding the new experience and the old one in mind together long enough to associate them. In other words, stress impaired the spontaneous retrieval of prior, related memories during new encoding. During stress, those old memory pairs couldn’t be brought up to help connect the dots between A and C.

Now for the boring parts that remind us that science isn’t a slam dunk (as much as the popular press wants a clean story to report). These were a hundred and twenty-one healthy young adults, screened to exclude anyone with an anxiety or mood disorder, exposed to a single sharp jolt of laboratory stress, with these associations being studied in a brain scanner, which is a long way from a person who has been marinating in worry for years, and then asked to defend their decision during a presentation in front of a large group of coworkers. Where does that leave us (and why would I bother to write about this study)? What we have is a real, mechanistic, carefully done study (a really careful and thoughtful design), but one that doesn’t directly connect the dots to the real world, especially that of mental health like anxiety, where people are constantly jacking their cortisol levels by worrying.

With those caveats in hand, the work still hands anyone who treats anxiety a real problem, and I want to flag it as a problem rather than dress it up as an answer. Anxiety is usually described in terms of overgeneralization. The classic finding is that anxious brains spread fear too widely, treating a safe thing as dangerous because it faintly resembles something that once wasn’t. That looks like the mirror image of what this study found, where stress made the brain keep things too separate. How do those two live in the same head? My best guess (and it is a guess, the kind I would want to test before I would want to assert it) is that they are running on different systems pulling in opposite directions at once: the fear circuitry smears threat across everything that rhymes with danger, while the hippocampal machinery that would let you integrate the reassuring counter-evidence (“I have felt precisely this and been fine”) goes quiet at the very same moment. The frightening associations generalize while the comforting ones fail to connect. If that is even partly right, it is a fairly cruel evolutionary hiccup, and it lines up uncomfortably well with what I watch happen in my clinic.

All of which circles back to the people I work with in my clinic. The thing we most want an anxious person to be able to do in the moment, to reach back across a hundred prior waves of anxiety and worry to feel, not just recite, that this one will pass too (or whatever they’ve been told to do to relax), is itself an act of memory integration, and it likely leans on the hippocampal weaving that stress appears to dampen right when the stress is running highest. Telling someone to remember they have done this before is asking the overwhelmed part of the brain to do the one thing that it has been blocked from doing. The span between what they already know and the moment they are standing in is a bridge too far.

In my own lab’s work, the practices that seem to actually help in those moments tend not to route through that kind of retrieval at all: dropping into the raw sensation in the body, getting curious about what the anxiety honestly feels like rather than what it threatens for the next twenty-four hours, all of it works in the present tense and does not need to build some cognitive bridge. I have wondered for years why “just remember last time” lands so flatly with people who remember last time just fine. This study leaves me suspecting that part of the answer is that, in the moment we most need it, the wiring that turns memory into useful inference is the wiring that builds that bridges is momentarily washed out with stress.

This is why I have my clinic patients and the members in my Going Beyond Anxiety program start with a grounding practice so that they can get their inference brain bits back online, BEFORE trying to take that next step. (See my previous articles on Five Finger Breathing, and The Simple Eye Trick for more). And this new study suggests why that might be so important: stress and anxiety keep us locked into very concrete thinking (and learning) patterns so that we (figuratively and literally) can’t open our minds to make connections or solve problems, let alone tap into our more creative sides.

The work that helps with stress and anxiety is less about handing people new information and more about keeping the lights on in the part of the brain that already holds everything they need to know.

Judson Brewer MD PhD is a psychiatrist and neuroscientist and professor at Brown University. He is the author of Unwinding Anxiety (NYTimes bestseller), The Craving Mind, The Hunger Habit and The Unwinding Anxiety Workbook. He co-founded MindshiftRecovery.org which provides free support for people with any type of addiction.

If you are struggling with anxiety, Dr. Brewer’s Going Beyond Anxiety program brings together his research and clinical experience to help people build effective skills to reduce anxiety and cultivate calm (www.goingbeyondanxiety.com).

Copyright © 2025, Judson Brewer, MD, PhD. All rights reserved.

REFERENCES

Armstrong, K., Avery, S., Blackford, J. U., Woodward, N., & Heckers, S. (2018). Impaired associative inference in the early stage of psychosis. Schizophrenia Research, 202, 86–90.

Brewer, J. A. (2019). Mindfulness training for addictions: Has neuroscience revealed a brain hack by which awareness subverts the addictive process? Current Opinion in Psychology, 28, 198–203. https://doi.org/10.1016/j.copsyc.2019.01.014

Cahill, L., Gorski, L., & Le, K. (2003). Enhanced human memory consolidation with post-learning stress: Interaction with the degree of arousal at encoding. Learning & Memory, 10, 270–274.

de Quervain, D. J.-F., Roozendaal, B., & McGaugh, J. L. (1998). Stress and glucocorticoids impair retrieval of long-term spatial memory. Nature, 394, 787–790.

de Quervain, D. J.-F., Roozendaal, B., Nitsch, R. M., McGaugh, J. L., & Hock, C. (2000). Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nature Neuroscience, 3, 313–314.

Kirschbaum, C., Pirke, K.-M., & Hellhammer, D. H. (1993). The “Trier Social Stress Test”: A tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology, 28, 76–81.

Laufer, O., Israeli, D., & Paz, R. (2016). Behavioral and neural mechanisms of overgeneralization in anxiety. Current Biology, 26, 713–722.

Lissek, S., Kaczkurkin, A. N., Rabin, S., Geraci, M., Pine, D. S., & Grillon, C. (2014). Generalized anxiety disorder is associated with overgeneralization of classically conditioned fear. Biological Psychiatry, 75, 909–915.

Madan, C. R., Scott, S. M. E., & Kensinger, E. A. (2019). Positive emotion enhances association-memory. Emotion, 19, 733–740.

Roozendaal, B., Okuda, S., de Quervain, D. J.-F., & McGaugh, J. L. (2006). Glucocorticoids interact with emotion-induced noradrenergic activation in influencing different memory functions. Neuroscience, 138, 901–910.

Schüren, K. A., Varga, N. L., Heinbockel, H., Preston, A. R., Roozendaal, B., & Schwabe, L. (2026). Stress disrupts hippocampal integration of overlapping events and memory inference in humans. Science Advances, 12(21), eaea5496. https://doi.org/10.1126/sciadv.aea5496

Schwabe, L., Hermans, E. J., Joëls, M., & Roozendaal, B. (2022). Mechanisms of memory under stress. Neuron, 110, 1450–1467.