ADHD working memory: why the thing you were holding disappears

In this article, ADHD working memory means the limited mental workspace used to hold information while doing something with it — like keeping the next step in mind while finishing the one before it — a working model for this article, not a clinical definition. Across adult studies, ADHD groups show measurable working-memory differences on average ¹,², but the pattern varies, and a larger pooled review found overall group performance was mostly at age-expected levels ¹⁰. So this isn't a rule that every ADHD brain has "bad memory," and it isn't a diagnostic test. For some, that gap can show up as an instruction or reason for opening a tab dropping out before it's used — one possible everyday shape, not a lab finding. This article looks at externalising that load — visible next steps, one capture point, less switching — as options to try, not proven fixes. Training studies show gains that tend to stay task-specific rather than transfer into daily life ⁶,⁷.

When ADHD and memory come up, this article uses a simple picture rather than a clinical account of how memory systems work: an open desk versus a filing cabinet — an invented image for this article, not a claim about what either system clinically is or contains.

In that picture, the desk is whatever's out and in use right now, for as long as it's needed — the working model for working memory described above. The cabinet is just a stand-in for everything else that isn't currently out on the desk. Neither is a claim that there's a literal container in your brain, or a technical account of how long-term memory works.

Here's one everyday shape it can take, as illustration rather than research: you're following a recipe on your phone, glance up to check the pan, and by the time you look back down you're not sure whether you already added the salt. The instruction hasn't vanished — it's still on the screen. What might have dropped is the "keep this active" part, whenever ADHD and memory come up — not a diagnosis and not unique to ADHD.

Researchers have also described working memory as one of several related skills — alongside planning and emotional self-regulation — within a broader executive-function group discussed in adult ADHD ³, covered more below.

None of this needs a lab to picture — what follows are invented, illustrative ADHD forgetfulness examples, not documented case studies and not a diagnosis. Imagine someone at work gives you two things to do — check the numbers, then send the update — and you're so deep in checking the numbers that the second half quietly slips. Or you open a new tab to look something up mid-task and, five seconds later, stare at the search bar with no memory of what you were about to type. Either way, the information hasn't vanished from the world — it's just no longer held in the part of the mind that keeps things active.

At home, a notification could pull attention away from making tea, so that coming back to the kitchen means losing the thread of what came next. In conversation, someone changing the subject for thirty seconds might make the point you were about to make evaporate, even though you know, abstractly, that you had one. And there's the version that can look like chaos from outside: holding three errands in mind on the way out, someone texts a fourth, and you come home having bought only the thing from the text.

None of these scenarios are evidence on their own — they're hypothetical scenes to help you picture what ADHD memory load might feel like, not a checklist and not a diagnosis. What the adult research does show is average group-level differences in working memory between people with and without ADHD, worth understanding on its own terms rather than read off a list of relatable moments ¹,²,⁹,¹⁰.

Here's what the evidence on ADHD and working memory in adults actually shows. A meta-analysis of 38 studies (combining results across many separate studies into one estimate) found moderate-magnitude group differences — real, but not enormous — across two domains: phonological working memory (verbal information) and visuospatial working memory (visual and spatial information) ¹. That's a real average pattern, not a claim that every adult with ADHD scores low.

A separate study using the WAIS-IV, a standard adult intelligence test, found moderate-to-large differences on subtests with working-memory and processing-speed demands, compared with matched controls aged 16 to 71 ². These are group averages from psychometric measures, not felt experiences of daily life.

Meanwhile, a much larger meta-analysis pooling over 1,800 people — the ADHD result a secondary finding inside a study mainly about autism, pooling children and adults together — found overall Wechsler test performance was mostly at age-expected levels, with only slightly reduced working-memory scores on average. The researchers were explicit that this profile isn't sensitive or specific enough for diagnosis ¹⁰. That also means this particular ADHD working memory test can't diagnose you off one score.

Worth naming honestly: a 23-to-25-year follow-up of just 19 adults with ADHD and 26 controls found the ADHD group scored below controls throughout on one working-memory task involving distraction, with a further decline later on ⁸. That's real, but from a very small sample on one narrow measure — nowhere near enough to support ADHD memory loss getting worse with age as an established pattern. One more point, a single snapshot rather than before-and-after: in a large young-adult cohort, lower working memory was associated with more inattention symptoms, though the link to hyperactivity and impulsivity depended on IQ ⁹.

Put together: across these studies, adult ADHD groups show a genuine average difference in working memory, but the size and pattern vary across studies and tasks. As the Wechsler-profile research specifically found, that profile isn't something you can use to test yourself for ADHD ¹⁰.

One account of adult ADHD's executive-function difficulties places working memory — both verbal and nonverbal — inside a broader group of mental jobs researchers call metacognition, meaning, plainly, the systems the brain uses to hold information, make a plan, and steer action towards it. The same account groups planning/problem-solving and emotional self-regulation alongside working memory under that broader heading ³. This is a narrative, theoretical account rather than a randomised trial or meta-analysis — a way of organising what researchers observe, not a proven mechanical fact about any individual brain.

Brain-imaging research adds a different kind of evidence. A meta-analysis combining 55 task-based fMRI studies (fMRI, functional MRI, is a type of brain scan) found ADHD linked to differences in activation across multiple neuronal systems, including one involved in working memory — extending earlier models that focused on one narrower brain circuit, rather than pointing to any single damaged region ⁴. Only 16 of the 55 studies were in adults, and the finding is a group-average, correlational pattern — useful for pushing back on a "one broken part" story, but not an individual diagnostic test.

Neither account proves what causes any one dropped intention on any one day. But together, they're a reason to treat a forgotten intention — the gap between deciding and acting — as a systems problem worth designing around, rather than a personal willpower failure.

It's a fair question: if you've searched ADHD working memory how to improve, is training the workspace itself the answer? There's a real evidence trail here, worth reading properly instead of trusting an app's marketing page.

A randomised controlled trial (participants are randomly assigned to a training group or a comparison group) tested Cogmed, a working-memory training programme, in adults with ADHD aged 18 to 63. The trained group did show greater improvement on measures of verbal and visuospatial working-memory capacity, and that improvement was still measurable six months later. But the gains did not carry over to nonverbal reasoning, day-to-day executive functioning, or ADHD symptoms more broadly ⁶. You get better at the drill. The drill doesn't reliably become your life.

That pattern isn't a one-off. A meta-analysis pooling 23 studies and 30 group comparisons of working-memory training — across clinical and typically-developing samples, children and adults together — found the same shape: short-term, task-specific improvement that doesn't reliably generalise to other cognitive skills, casting real doubt on how clinically useful these programmes are ⁷. And in a small trial scanning the brains of adults with ADHD before and after five weeks of training, the training produced no measurable change in functional or structural brain connectivity versus an active control group ⁵.

None of this means training is worthless or that nothing helps. It means the specific evidence gathered here doesn't support training your way to a bigger working memory that then fixes daily life on its own. So rather than promising to add capacity, the rest of this article focuses on reducing how much the workspace has to carry — a design choice worth trying, not a proven treatment.

Because working-memory training does not reliably transfer into daily life, one option is designing the environment to carry more of the load. Four jobs matter here, offered as things to try — not a validated system, and deliberately not a 47-item productivity list.

Capture it. Have one immediate landing place for instructions, thoughts, and commitments the moment they arrive — a single notes app, a single notebook, whatever you'll actually reach for mid-sentence. This isn't a tested clinical strategy; it's an idea worth trying if you're often unsure where things went once thought.

Show the next step. One idea worth testing for yourself: instead of facing an undifferentiated pile of "stuff to do," turn it into one visible next action — not the whole plan, just the next physical step. A practical suggestion, not a demonstrated effect.

Reduce switching. Keeping whatever you're working on and the materials it needs together, physically or digitally, is another option to consider — the idea being that returning after an interruption doesn't mean rebuilding context from nothing. Untested in the research above; a design suggestion, not a proven outcome.

Recover the system. Capture apps, lists, and reminder setups can go unused for a stretch — not necessarily a personal failing, and maybe just what happens to systems sometimes. Building in a way back in before you need it is one option worth having, not a guaranteed fix, so restarting doesn't become a task big enough to avoid on its own.

James, describing one habit that works for him, put it simply: working with AI tools has helped him because their limited scope forces him to state upfront what he's actually working on — "this is what I'm working on" — which acts as a kind of externalised task definition. That's not evidence that AI tools improve working memory. It's one person's example of the same underlying move: get the thing out of your head and into something you can see ¹¹.

If the problem you're hitting has shifted from holding the next step to actually starting and sticking with the task once it's visible, ADHD productivity systems that work with your attention is the place to look instead. And if what you're describing isn't one dropped thought but your whole system feeling flooded at once, when working-memory load becomes ADHD overwhelm is a separate topic worth reading on its own.

Reminder apps can get deleted. Notebooks get lost. A list system can quietly stop being opened. None of that means you failed — it can just mean the system needs a way back in, not only a way to start.

A few things worth building in from the start, offered as options rather than rules:

• One capture point, not three competing ones, so restarting doesn't mean choosing between half-finished systems first.
• Fewer decisions at restart — if picking the system back up requires reorganising everything before you can use it, that's a barrier disguised as tidiness.
• One visible re-entry cue — something that shows you where you left off without requiring you to remember, like a pinned note or an open tab, rather than a closed app you have to remember exists at all.
• A periodic, low-pressure check on whether the system is still reducing what you're carrying, or has quietly become one more thing to maintain. If it's the latter, that's information, not failure.

None of this is a clinically validated treatment or a cure. It's a set of design options to try — not a description of how ADHD works, and not a guarantee that any of them will fit you.

You don't need a whole new system to start. Pick the one thing you lose most often — the second half of instructions, the reason you opened the tab, whatever it is for you — and try giving just that one thing a single visible landing place, as a small experiment rather than a fix. Before you add anything else, try making its very next step visible too, rather than leaving it buried in a plan.

If the gap you're noticing is more about starting than holding, ADHD productivity systems that work with your attention picks up from here. If it's less one dropped thought and more everything at once, when working-memory load becomes ADHD overwhelm is the next place to look.

¹ Alderson RM, Kasper LJ, Hudec KL, Patros CH. Attention-deficit/hyperactivity disorder (ADHD) and working memory in adults: a meta-analytic review. Neuropsychology. 2013 May;27(3):287-302. https://doi.org/10.1037/a0032371. PMID 23688211.

² Theiling J, Petermann F. Neuropsychological Profiles on the WAIS-IV of Adults With ADHD. Journal of attention disorders. 2016 Nov;20(11):913-924. https://doi.org/10.1177/1087054713518241. PMID 24448224.

³ Barkley RA. Differential diagnosis of adults with ADHD: the role of executive function and self-regulation. The Journal of clinical psychiatry. 2010 Jul;71(7):e17. https://doi.org/10.4088/JCP.9066tx1c. PMID 20667287.

⁴ Cortese S, Kelly C, Chabernaud C, Proal E, Di Martino A, Milham MP, Castellanos FX. Toward systems neuroscience of ADHD: a meta-analysis of 55 fMRI studies. The American journal of psychiatry. 2012 Oct;169(10):1038-55. https://doi.org/10.1176/appi.ajp.2012.11101521. PMID 22983386.

⁵ Tolonen T, Leppämäki S, Roine T, Alho K, Tani P, Koski A, Laine M, Salmi J. Working memory related functional connectivity in adult ADHD and its amenability to training: A randomized controlled trial. NeuroImage. Clinical. 2024;44:103696. https://doi.org/10.1016/j.nicl.2024.103696. PMID 39536524.

⁶ Dentz A, Guay MC, Parent V, Romo L. Working Memory Training for Adults With ADHD. Journal of attention disorders. 2020 Apr;24(6):918-927. https://doi.org/10.1177/1087054717723987. PMID 28856932.

⁷ Melby-Lervåg M, Hulme C. Is working memory training effective? A meta-analytic review. Developmental psychology. 2013 Feb;49(2):270-91. https://doi.org/10.1037/a0028228. PMID 22612437.

⁸ Torgalsbøen BR, Zeiner P, Øie MG. Pre-attention and Working Memory in ADHD: A 25-Year Follow-Up Study. Journal of attention disorders. 2021 May;25(7):895-905. https://doi.org/10.1177/1087054719879491. PMID 31625421.

⁹ Soares PSM, de Oliveira PD, Wehrmeister FC, Menezes AMB, Rohde LA, Gonçalves H. Does IQ Influence Association Between Working Memory and ADHD Symptoms in Young Adults? Journal of attention disorders. 2022 Jun;26(8):1097-1105. https://doi.org/10.1177/10870547211058813. PMID 34779307.

¹⁰ Wilson AC. Cognitive Profile in Autism and ADHD: A Meta-Analysis of Performance on the WAIS-IV and WISC-V. Archives of clinical neuropsychology : the official journal of the National Academy of Neuropsychologists. 2024 May 21;39(4):498-515. https://doi.org/10.1093/arclin/acad073. PMID 37779387.

¹¹ James. Personal lived-experience extract for this article, 30 July 2026. This is one person's example and is not evidence that AI tools improve working memory.