ADHD and Working Memory: Why Things Fall Out of Your Head

Working memory is the small mental workspace that holds a few things active while you are using them. In ADHD that workspace holds less and lets go sooner, which is why an intention can disappear between the sofa and the kitchen. It is one of the most consistently measured differences in the ADHD research, and one of the most workable, because the answer is a better place to put things rather than a better memory.

Five versions of the same moment:

Working memory is smaller than an ordinary day expects

Working memory is not a synonym for memory in general. Alan Baddeley and Graham Hitch described it in 1974 as a set of parts working together rather than one storage box: a loop that holds words and sounds, a sketchpad that holds shapes and places, and a central executive that decides where attention goes and keeps the whole thing coordinated. Baddeley later added an episodic buffer, the piece that binds a name, a face, and a room into one connected scene (Baddeley, 2000).

It is small in everyone. The popular figure is seven items, from George Miller’s 1956 paper on the “magical number seven” (Miller, 1956). Miller was describing chunks rather than raw objects, and a familiar year like 1492 counts as one chunk instead of four. Under tighter conditions that prevent rehearsal and chunking, the more defensible estimate is roughly four (Cowan, 2001). Current work treats three to five as the honest range.

It is also short. The textbook rule of thumb is that unrehearsed material stays usable for something like fifteen to thirty seconds. Treat that as a rough heuristic rather than a fixed limit, because researchers increasingly think interference from competing material explains more of the loss than the passing of time does. Either way, the everyday feeling of losing a thought while somebody talks over you is the system behaving as described, and it says nothing about how much you care.

What ADHD does to it

Adults show the difference too, at moderate size. A meta-analytic review of 38 adult studies found impairments in both the verbal and the visual-spatial kinds, at roughly g = 0.55 and g = 0.49, which is squarely in the moderate range (Alderson, Kasper, Hudec & Patros, 2013). Those are adult samples measured against adult controls rather than the same people followed from childhood, so the fair reading is that the difficulty is still present in adulthood, not that it travels forward unchanged.

In children the visual-spatial side shows the larger gap. A meta-analysis of 26 studies found spatial storage effects around 0.85 and spatial executive effects around 1.06, against verbal effects near 0.43 to 0.47, and those gaps held after accounting for language-learning disorders and general intellectual ability (Martinussen, Hayden, Hogg-Johnson & Tannock, 2005).

In one closely modeled sample, the manipulation part was much harder than the holding part. Keeping four items still is one job. Keeping them while reordering them, comparing them, or ignoring something noisy is a different and much harder one. A 2020 study of school-age children using bifactor modeling found very large impairments in central executive working memory, d = 1.63 to 2.03, with 75 to 81 percent of the ADHD group classified as impaired. The same study found a smaller visual-spatial short-term memory effect of d = 0.60, at 38 percent impaired, and no evidence of a phonological short-term memory deficit at all (Kofler et al., 2020).

That pattern is the useful one, with the caveat that it comes from a single school-age sample and one modeling approach. If it holds more widely, the difficulty sits in holding things while doing something with them, which is what almost every real task asks of you.

And it is not universal. In that same study, 19 to 25 percent of the children with ADHD were not classified as impaired on the executive measures. Working memory difficulty is common in ADHD, it is not required for the diagnosis, and no test result of any kind diagnoses ADHD on its own. Some people with ADHD struggle far more with inhibition, timing, or getting started than with holding things in mind.

Why it costs more than a few forgotten items

The workspace is upstream of most of the things that go wrong in a day. Following a three-step spoken instruction needs steps two and three held while you do step one. Reading needs the earlier sentences held while your eyes take the next. Mental arithmetic, a conversation with two threads, cooking a meal with three timings, deciding what to do next out of eleven options: all of it runs through the same small space.

It also stacks with the other parts of ADHD. A 2024 systematic review and meta-analysis pooled 824 effect sizes and found the size of the time-perception deficit in ADHD tracked with working memory (Metcalfe, McFeaters & Voyer, 2024), which is part of why “I will deal with it in an hour” is such an unreliable plan, and why ADHD time blindness rarely travels alone. When an intention cannot stay active without an outside cue, things that leave your sight leave your plans, which is the mechanism underneath what people call ADHD object permanence.

What an ordinary Tuesday looks like

• You open a browser tab to look something up and find yourself reading something else, with no memory of the original question.
• Somebody gives you their address out loud and it is gone before your thumb reaches the notes app.
• You start unloading the dishwasher, spot a mug that belongs upstairs, and come back down having made the bed and not finished the dishwasher.
• You put a saucepan on, sit down for two minutes, and remember it when it starts to catch.
• You have a good idea in the shower and it is fully gone by the time you have a towel.

None of that is a character flaw, and effort is not the missing ingredient. What helps is leaving yourself better evidence.

The move that works: put it outside your head

Russell Barkley, whose 1997 paper set out the standard executive-function account of ADHD, reached the practical conclusion decades ago. When information cannot be reliably held inside, put it outside. In his words, “to externalize information is to make it physical outside of the individual,” and those physical cues need to sit at the point of performance, meaning the place and the moment the action has to happen (Barkley, ADHD, EF and SR fact sheet).

The laboratory evidence backs the principle, with a couple of honest edges.

Saving something to a place you trust improves what you learn next. In three experiments, people who saved a first list of words to a computer before studying a second list remembered the second list better than people who did not. The effect depended on the saving being reliable. If the storage could not be trusted, the benefit disappeared (Storm & Stone, 2015). That was word lists in a laboratory rather than errands in a kitchen, so treat it as evidence about trusted storage rather than about any particular app.

Offloading changes the demands of the task rather than expanding your capacity. The review that defines the field describes cognitive offloading as using physical action to reduce the mental demand of a task, covering everything from calendars and sticky notes to counting on your fingers. It also names the cost plainly: leaning on an external aid can reduce your unaided memory for what you offloaded (Risko & Gilbert, 2016). The review does not settle whether that trade is worth making. It does argue for keeping one system rather than seven, since the cost lands on whatever you hand off and stop rehearsing.

A specific plan quiets an unfinished task. Across five studies, unfinished goals produced intrusive thoughts during unrelated reading and measurably hurt performance on unrelated puzzles. Asking people to make a concrete plan for the unfinished goal removed those effects, without the goal being completed (Masicampo & Baumeister, 2011). The plans that worked were specific about when and where. “Email Alex the draft at nine, at my desk” does the job. “Deal with the draft” does not. What was compared was a specific plan against no plan, rather than a plan against an ordinary list, so the supported claim is about specificity.

This is the reasoning behind a brain dump, and behind the small step that has to follow it. Getting everything out is the relief. Deciding when the one thing happens is what stops it circling back.

Supports worth building

CHADD’s practical guidance for adults comes down to the same principle applied in a lot of small places (CHADD, Make Memories with ADHD):

1. One capture place, not five. Groceries, appointments, ideas, and tasks all land in the same trusted spot. Five lists behave like no list.
2. Notes instead of retention in meetings. Write it or record it. Holding a spoken plan while nodding along is the exact operation that is hardest.
3. Checklists for anything with more than three steps. The steps you do every week are still steps you can lose (CHADD, Workplace Issues).
4. Cues where the action happens. The pill bottle by the kettle, the gym bag against the front door. A reminder in the wrong room is a reminder you will not see.
5. Capture before you switch. When something interrupts you, write down where you were before you deal with it. Coming back to a note is fast. Reconstructing a lost thought is not.
6. One thing at a time, with a place to put the rest. Keep a scrap page beside you for the distracting ideas so they do not have to be held or chased (CHADD, Workplace Issues).

If getting started is the harder half for you, task initiation covers that part, and ADHD organization tips covers the environment side.

Can you train working memory?

Mostly not in a way that shows up in your life, and this is worth knowing before you pay for anything.

A meta-analysis of 87 publications and 145 comparisons found that working memory training reliably improves performance on the trained exercises and on very similar tasks, and produces no convincing improvement in intelligence, verbal ability, reading comprehension, or arithmetic when measured against active control groups. The amount of gain on the training tasks did not predict the size of any wider benefit (Melby-Lervåg, Redick & Hulme, 2016). A large review of the commercial brain-training field reached the same conclusion: better scores on the game, little evidence of better everyday cognition (Simons et al., 2016). Both are general reviews of working memory and brain training rather than ADHD-specific ones, so read them as evidence about the training itself.

So getting better at the app is not the same as remembering the laundry. The supports above are the boring answer. Nobody has run the head-to-head trial, so the honest framing stays narrow: training has been tested repeatedly and keeps failing to transfer, while the supports rest on the offloading work above.

What this does and does not claim

Three limits, because this topic collects confident claims.

Every effect size here is a group average, and group averages describe distributions rather than people. A meaningful minority of people with ADHD test in the typical range on working memory measures, and that does not make their ADHD less real.

The offloading studies are laboratory studies with word lists and puzzles, run mostly on general populations rather than on adults with ADHD. The principle is well supported and the mechanism is plausible. The specific claim that a to-do list improves ADHD outcomes has not been tested that directly.

And none of this is medical advice. ADHD is treatable, treatment is a conversation with a clinician, and a planner is not treatment. If forgetting is costing you your work, your health, or your relationships, that is worth raising with a doctor rather than solving with an app.

Where Whit fits

The method matters more than the tool. A pocket notebook you always carry beats any app you do not open, and everything above works with paper.

What a good tool changes is the trust and the follow-through. The offloading effect depended on the storage being reliable, and reliability is a real design problem. Whit is built around Brain Dump capture by text or voice, so a thought reaches a safe place in the seconds before it goes, and everything lands in one Calm Inbox instead of scattering across five apps. The Visual Day is where the second half happens: instead of a list of things you owe, the day shows as a gentle timeline from sunrise to night, so “when does this actually happen” gets an answer, which is the part the plan-making research says does the quieting.

There are no red overdue badges and no streak pressure, because a tool that shames you for a bad week is one more thing to avoid. Whit syncs only through your own iCloud, with no accounts, no logins, and no analytics in the app. You can see the rest of the features if you want the detail.

Common questions about ADHD and working memory

Is working memory always impaired in ADHD?

No. Group differences are consistent and moderate to large, but a 2020 study found 19 to 25 percent of children with ADHD were not classified as impaired on central executive measures (Kofler et al., 2020). Working memory difficulty is a common associated feature, not a diagnostic requirement, and no cognitive test diagnoses ADHD by itself.

How many things can working memory hold?

Fewer than most people assume. The familiar figure of seven comes from Miller in 1956 and refers to chunks rather than single items (Miller, 1956). Under conditions that prevent rehearsal and chunking, the better estimate is about four, with three to five as the honest range (Cowan, 2001).

Why do I forget what someone just said to me?

Because spoken information arrives faster than it can be stored, and holding it while also listening for your turn draws on the same limited space. In ADHD the manipulation part looks like the harder half. Asking for the message in writing, or repeating it back as you hear it, moves the work outside your head where it is safer. Remembering to act on it later is a separate job, covered in ADHD forgetfulness.

Does brain training improve ADHD working memory?

Not in ways that show up in daily life, though the evidence is general rather than ADHD-specific. A meta-analysis of 87 publications found training improves the trained tasks with no convincing transfer to intelligence, reading, or arithmetic (Melby-Lervåg, Redick & Hulme, 2016). No trial has compared training against everyday supports.

Is a brain dump a working memory strategy?

In effect, yes. A brain dump is cognitive offloading applied to everything at once: moving open loops out of a small internal workspace into a reliable external one. The underlying studies used word lists and single goals rather than brain dumps, and what they favor is trusted storage plus specific follow-up (Masicampo & Baumeister, 2011).

Related reading

• The capture habit this all points to: what a brain dump is.
• Turning the pile into an actual day: from brain dump to daily plan.
• Why later is such a shaky plan: ADHD time blindness.
• When things leave your sight and your plans: ADHD and object permanence.
• Getting from knowing to starting: task initiation.
• The invisible tracking that fills the workspace: the mental load.

Where to start

Pick the moment that costs you most this week. The shower idea, the spoken address, the interrupted task. Then change one thing so that moment has somewhere to land: a note by the door, a voice memo before you dry off, three words written before you stand up.

Holding more was never on offer. Somewhere you trust to put things down, plus a plan specific enough to answer when, does the job that memory keeps failing to do.

If you want that place to be calm, Whit is a one-time purchase for iPhone, iPad, and Mac. There is a 7-day free trial, then $99 once with no subscription. See the details.