Skip to main content
Kioku Games

N-back

Press MATCH when the current symbol equals the one N steps back.

N-back

How to Play: N-back

🎯 Goal

Press MATCH when the current symbol equals the symbol shown N steps ago.

🖱️ Mouse / Touch

  • Watch the symbols flash in the center
  • Click MATCH (or press Space) when current = N back
  • Wait through if it doesn't match — no penalty for waiting

⌨️ Keyboard

Space / Enter Press MATCH instantly

💡 Tips

  • Sub-vocalize: say "cat-dog-cat" → match
  • Most adults max out around N=3

More games

See all →

About this game

N-back is the most-studied working-memory task in cognitive science. The rule is simple but the demand is unforgiving: a stream of symbols flashes one at a time in the centre of the screen, and your job is to press MATCH whenever the current symbol equals the one shown N steps earlier. At N=1 you compare the current symbol to the previous one — easy. At N=2 you compare it to the one before the previous one — already noticeably harder. At N=3 most adults start to falter, and N=4 separates serious practitioners from everyone else. Each round runs 20 symbols.

What makes N-back so brutal is that you cannot pause to reflect. While you decide whether the current symbol matches "N back", a new symbol is already arriving — and it will become the new "N back" target a few moments later. You are continuously updating a sliding window of items in working memory, replacing the oldest with the freshest, all while making a comparison decision on every step. This double duty (storage + manipulation) is exactly why cognitive psychologists treat N-back as the cleanest available probe of executive working memory.

The cognitive science of N-back

Working memory is the system that lets you hold and manipulate information for short periods — phone numbers, conversation context, the goal of a multi-step task. Alan Baddeley's influential model splits it into a phonological loop (for verbal info), a visuospatial sketchpad (for visual/spatial info), an episodic buffer (for binding), and a central executive (for control). N-back hammers the central executive in particular: every step requires inhibition (suppressing irrelevant past items), updating (replacing the oldest item with the newest), and shifting (between maintaining the window and producing a response). These three are sometimes called Miyake's "executive functions trinity".

Neuroimaging studies consistently show N-back activates the dorsolateral prefrontal cortex, the parietal cortex, and the anterior cingulate — the same network that lights up when you do mental arithmetic, plan a multi-step project, or hold a complex conversation. As N increases from 1 to 3, brain activation in these regions increases roughly linearly, which is why N-back is often used in fMRI studies as a parametric working-memory load probe. The takeaway: when you play N-back, you are exercising precisely the circuitry your brain uses for the most cognitively demanding parts of daily life.

A short history: Kirchner, Jaeggi, and the IQ debate

The N-back task was designed by Wayne Kirchner in 1958, originally to compare working-memory function in young versus older adults. He had subjects watch a sequence of lights and press whenever the current light matched the one N steps back. The task stayed mostly inside cognitive psychology labs for half a century, used as a clean parametric tool but not exactly a household name. That changed in 2008.

In 2008, Susanne Jaeggi and colleagues published a study in PNAS reporting that 19 days of "dual" N-back training produced measurable gains on tests of fluid intelligence (Gf) — the kind of reasoning that supposedly underlies general IQ. The internet exploded. For a few years, dual N-back was treated as a one-weird-trick to raise IQ, with thousands of self-taught practitioners building spreadsheets of their progress. The follow-up evidence has been more sober: meta-analyses (Melby-Lervåg & Hulme, Sala & Gobet, others) show robust improvements on the trained task itself but small, inconsistent transfer to broader IQ. The honest summary is that N-back makes you better at N-back and at very similar tasks, with modest near-transfer and questionable far-transfer. That is still useful — your working memory for symbol streams genuinely improves — but it is not an IQ cheat code.

Strategies that push your N

Three habits separate competent N-back players from grinders who plateau at N=2. First, sub-vocalise the symbols as a rhythmic chant. If you see "cat — dog — cat", inwardly speak it as a three-beat phrase. The phonological loop is your friend at low N values; verbal labels make the comparison much faster than purely visual matching. The chant also forces a steady internal tempo, which suppresses the panic that builds when symbols come fast.

Second, do not "remember" — let the recent items live in working memory naturally and just compare. Players who try to deliberately hold the last 3 items in mind exhaust the central executive within a round and crash. Players who treat each step as "what was N ago, vs what is now" let the brain's natural decay-replacement dynamics handle the storage, and only intervene at the comparison. Third, accept errors. The signature of strong N-back play is not zero false alarms — it is calm recovery after each error. Tilting after one mistake costs you three more.

Difficulty, scoring, and how Kioku N-back works

Each round runs 20 symbols at a steady pace; you choose your N before the round starts (N=1 through N=4). The game scores hits (correct MATCH presses) and false alarms (incorrect MATCH presses) separately, with the headline score being d-prime — a signal-detection metric that rewards hits and penalises false alarms in a single number. d-prime is the same metric used in published N-back research and avoids the trap of "spam-press MATCH and look perfect on hits". A round of 18 hits / 0 false alarms beats 20 hits / 5 false alarms, every time.

Personal best is saved per N value in your browser local storage, so you can track your N=2 best independently of your N=3 best. The optional weekly leaderboard accepts the highest d-prime score you posted at any N (with N tagged); reset is every Monday at 00:00 UTC. There is no time pressure between rounds and no penalty for waiting through a round to settle nerves before the next one.

N-back for students, professionals, and seniors

For students preparing for exams, N-back is one of the few brain-training tasks with credible (if modest) transfer to academic outcomes — specifically, to working-memory-heavy operations like multi-step math, reading comprehension across long sentences, and following instructional sequences. Five to ten minutes a day for two or three weeks is the protocol most often cited in studies. Going harder than that produces fatigue without proportional benefit; the dose-response curve flattens quickly.

For professionals in cognitively demanding fields (programmers, surgeons, traders, simultaneous interpreters), N=3 daily is a useful "mental warm-up" — it reliably puts you into a state of focused executive engagement that carries over for the next hour or two. For older adults, the value is different: studies of cognitive aging consistently find that working-memory training slows the gentle decline of executive function with age, and N-back is the most-validated paradigm for this purpose. N=1 daily is a reasonable starting point at any age past 60; the goal is consistency, not heroic N values.

Related Kioku Games

If you enjoy N-back, three sister games extend the working-memory theme. Digit Span is the classical span measure — store a sequence and reproduce it; cleaner storage, less continuous updating. Quick Math layers arithmetic on top of working-memory storage, requiring you to hold operands while computing. Color Sequence pivots the same kind of memory load entirely onto the visuospatial sketchpad with no verbal route. Together these four cover storage, manipulation, and updating across both verbal and visual channels — a surprisingly complete working-memory workout in twenty minutes a day.

Frequently asked questions

Is N-back really free?

Yes, completely free. No signup, no in-app purchase, no premium tier. The site is supported by display ads next to the game.

Do I need to create an account?

No. Personal best per N value is stored locally in your browser. You only need to enter a display name if you choose to submit a score to the optional weekly leaderboard.

Will N-back raise my IQ?

Probably not by much. The original 2008 Jaeggi study made that claim, but follow-up meta-analyses show consistent improvements only on the trained task and very similar tasks (near transfer), with small or inconsistent transfer to general IQ measures (far transfer). The honest claim: your working memory for symbol streams improves reliably; broader cognitive gains are uncertain.

What's the difference between this and "dual N-back"?

Dual N-back tracks two streams simultaneously — typically one visual (position on a grid) and one auditory (spoken letter), with separate match buttons for each. It is harder by design. Kioku N-back is the single visual variant, which is gentler, equally well-studied, and easier to fit into a 5-minute session.

What N is "good"?

N=1 is the warm-up; almost anyone can hit good d-prime there. N=2 is the realistic working level for most adults; expect to plateau there for a few sessions. N=3 separates focused practitioners; reaching consistent positive d-prime at N=3 takes weeks of practice. N=4 is uncommon and impressive.

Why does it feel so much harder than digit span?

Digit span tests storage. N-back tests storage plus continuous updating plus comparison, all at once, under time pressure. Three executive functions instead of one. The added load is exactly what makes it the cleanest probe of executive working memory.

How long should I practice?

5-10 minutes per day is the dose used in most published training studies. Going past 20 minutes per session produces fatigue without proportional benefit. Consistency matters more than session length.

Can I play with a keyboard?

Yes. Space or Enter triggers MATCH. The whole game is fully keyboard playable, and most experienced players prefer it over clicking — the response is faster and lets you focus on the decision rather than aiming.

Is it OK for kids?

Yes from about age 9, with N=1 only. The task requires sustained attention that younger children struggle to maintain for a full round. There is no chat, no user-generated content, and no in-app purchases.

How does the weekly leaderboard work?

Submitting your highest d-prime score is optional and only requires a display name (no email). Scores are tagged with the N value used. The leaderboard resets every Monday at 00:00 UTC, and the previous week's top is archived as "Hall of Fame".