Published in Behavioral and Brain Sciences · The finding of about four units
The name you have lost, the word that drops away, that "what did I come here for again". Researchers have worked out how much a person can hold at once, and that number is small. A free five minute check shows you how your clarity is doing.
"I used to keep everything in my head. Now I write everything down and I still lose things."
If that sentence is true for you, read on.
I am Jesse van der Velde. Not a guru and not a doctor. I study and share what published research says about how human memory really works.
I want to give you one number. It changes what you say to yourself when the next name does not come up.
The number is four. Not forty. Four.
VIDEO SLOT · place a video or embed here
When a word drops away halfway through your sentence, the verdict is there at once. And past 45 a story comes with it: it is starting, I am going downhill, my memory gets worse every year. At birthdays we make jokes about it that are too hard.
But you have never measured that.
With that judgement you compare yourself with how you think you used to be. That picture comes from your memory.
What has changed is how much comes at you per hour. More channels. Switching more often. More things open at the same time.
The wrong enemy is your brain. And below I will show you why.
In 2001, Behavioral and Brain Sciences published an analysis by Nelson Cowan.¹
What was studied: how many separate things a person can hold in working memory at once. That is the memory you use while you are busy with something.
How: Cowan laid the capacity experiments from decades of research side by side and looked at four kinds of situation in which that limit becomes visible: when so much information arrives at once that grouping fails, when grouping and silent rehearsal are blocked, when performance drops off at a certain point, and where that same limit works through indirectly.
The findings: under those conditions the reliable capacity lies around four units on average. Researchers call such a unit a "chunk": one lump of information that you hold as a single thing. Others had already arrived at three to five before that, and Cowan argues that this smaller limit is the real one. The number seven from 1956 was, in Cowan's view, more a rough estimate and a rhetorical device than a real limit.
Put plainly: you hold about four at once.
And be honest about what this piece is. Cowan ran no experiment. The article appeared together with open commentary from peers, a form of publication built on contradiction. The four is an average from tightly defined laboratory tasks, not a fixed property of your head and not a measure of your memory in daily life.
Add up what your Tuesday wants to put on those four: the meeting, the unanswered message, picking up the children, and that thing you walked into the kitchen for.
There is something more on top of that.
In 2003, Trends in Cognitive Sciences published a review article by Stephen Monsell on task switching.²
What was studied: what happens when you move from one task to another.
How: Monsell summarised experiments in which people had to keep switching between a small number of simple tasks, so that the cost of the switch itself can be worked out. No new data was gathered for it.
The findings: straight after a switch, people are clearly slower and usually make more mistakes. Researchers call that the "switch costs".
In some of the experiments people knew in advance that the switch was coming. That is the comparison the first study does not have. The switch costs then become smaller. They do not go away.
Monsell points to two causes: the previous task carries on in your head, sometimes briefly and sometimes much longer, and resetting your attention to the next task takes time.
What this study cannot tell you: it is about simple lab tasks from 2003, with no notifications, email or phone. It gives no seconds and no percentages. Anyone who tells you how many minutes an interruption costs is taking that number from somewhere else.
Then there is the state you do all of this in. That has been measured too.
In 2026, Medicine and Science in Sports and Exercise published a randomised crossover study with 117 adults.³
What was studied: what prolonged, effortful thinking does to reaction time on a GoNoGo test.
How: 117 participants, 57 of them women, average age 32. Everyone did both sessions. One time 45 minutes of a Stroop task, matched to the participant's level. The other time a documentary of their own choosing, just as long. After each one, the same six-minute GoNoGo test, in which you alternately have to respond or hold your response back.
The findings: after the Stroop task, the reaction time on the GoNoGo test was worse than after the documentary (P < 0.001).
This setup can do something the previous two cannot: the same people did both sessions, so you compare someone with themselves. What then appears is called "mental fatigue": not a mood, but something that shows up in the measurements.
What it cannot tell you: that fatigue was brought on with a lab task; a real working day full of screens was not tested. The measured effect is small (ηp2 = 0.07). No blinding is described. And neither nutrition nor a supplement was tested in it.
Put the three findings side by side.
Forty is not a measurement. It is the order of magnitude of an ordinary day, and that sits far above four.
None of these three measurements says that there is anything wrong with your head.
Does this sound familiar? Take the free check now.
Our scientifically validated questionnaire gives you answers
About 4. The number of units your working memory reliably holds at once, averaged over decades of experiments.¹
Every switch costs you something. Straight after a task switch, people are slower and usually make more mistakes. Announcing it beforehand makes those costs smaller, but does not take them away.²
45 minutes. That is how long the Stroop task lasted in the 2026 crossover study. Afterwards, the reaction time of 117 adults was worse than after the documentary.³ Count the 45-minute blocks in your day.
You introduce two people to each other. You know both names. One of the two you have known for eleven years.
And right at the moment you want to say the name, it does not come. You keep smiling while the colour rises in your face.
Look at what happens then. You hold about four at once,¹ and you are already busy with the introduction itself, with who is listening in and with what you wanted to say next. You have been switching between tasks all morning, and every switch cost you speed and accuracy.² And you are doing this with a head that had already been thinking hard for hours that morning, which makes the response measurably slower.³
Twenty seconds later the name came after all. As it always does.
The problem was never your head. It did what a head does. The problem was that you did not know how much it holds at once.
Think about this for a moment. None of those three measurements is about decline.
Writing everything down. Sensible, and please do it. But be clear about what it does and what it does not do. A list takes the storage out of your head: you no longer have to remember that something needs to happen. What it does not do: in the moment itself you still hold only about four.¹ And checking seventeen lists means switching seventeen times, with those costs again at every switch.² You have put your storage in order and left your working memory just as full.
The brain training apps. How you spend your free time is of course up to you. But be honest about what you are buying: puzzles make you better at puzzles. Your problem is how much you hold during a real day, and in what state you do it. An app measures neither, and the subscription keeps running.
Laughing it off. "Old age, haha." It works for a moment: you get through the moment and nobody asks further. But be clear about what it costs you. Out loud, you give yourself an explanation that rests on no measurement at all. Nothing changes.
Three strategies, one shared problem: none of them looks at how much you can hold, and none of them measures the state you do it in.
Your head holds about four at once, and your day asks for many more. That difference is what you have come to call decline.
It is not you falling short. The strategies fall short, and you chose them without ever knowing where you stood. It can be different.
This is the essence. Once you know how much your head holds at once and what each switch costs, the work falls apart into two steps.
The condition for step two: the analysis has to come from your own pattern. From your days, your symptoms, your context. Not from a birthday joke or one bad Tuesday.
And the order is fixed. Working on your state without first knowing where you stand is guessing again, and you have made that mistake often enough.
You are five minutes away from that analysis, and it is free.
BioProfile has built that analysis. A scientifically validated questionnaire, five minutes, free and with no account. It maps out how your focus, your clarity and your energy are doing, and shows where your answers deviate most strongly. Not a medical examination. An analysis of your own answers. You use the outcome yourself, or you put it on the table with your doctor or your coach.
And then the daily side. I do that the same way every morning: Green Juice, a plant-based all-in-one daily drink mix with more than 44 ingredients, including vitamins, minerals, herbs, grasses, fruit, vegetables, algae, probiotics and digestive enzymes. The B vitamins in it and minerals such as chromium and potassium contribute to a normal energy-yielding metabolism and to the reduction of tiredness and fatigue through the day.
To be precise: it is a nutritional base. It makes no promise about your memory and none about your concentration. The analysis tells you where you stand now. How often you switch tasks per hour stays your choice.
The next time a name does not come, you know what has been measured. About four units at once.¹ Switch costs every time you switch task.² And the state you are in at that moment.³
Then you do not immediately conclude that you are going downhill.
What is still missing is the analysis itself. You have been judging yourself for years without ever having measured where you stand. Do it now: five minutes, free, no account, and there is nothing to buy. Then read the outcome.
Our scientifically validated questionnaire gives you answers
The scan is not a medical examination. It is an analysis of your own answers. If complaints persist, consult your doctor.
