What Is the N-Back Task?
The N-back is probably the most-used working-memory task in cognitive research. The rules are simple to state and genuinely hard to do: a stream of stimuli appears one at a time — letters, positions on a grid, sounds — and for each one you decide whether it matches the item presented N positions back. At 2-back you compare against two items ago; at 3-back, three.
The task entered the mainstream after Jaeggi and colleagues' 2008 study reported that practicing adaptive N-back transferred to fluid intelligence. That claim mostly failed to replicate — later work and meta-analyses found training gains stay close to the practiced task (Simons et al. 2016; Sala & Gobet). But the N-back itself survived the controversy as an instrument: it's still a standard way to load working memory in the lab, and a real community trains on it for the challenge itself.
What makes a real N-back implementation
A few details separate an honest N-back from a lookalike:
- Controlled match frequency. Matches should be planted at a fixed quota (roughly a third of trials is common), not left to chance — the mix of matches and non-matches is what forces you to update the window every single trial instead of pattern-matching.
- Lures. An item that matches the position adjacent to N back (an "n−1 lure") is the classic source of false alarms. Implementations that include lures discriminate between real updating and lucky guessing.
- Multiple streams. The original Jaeggi task was dual N-back: a spatial stream and an auditory stream judged simultaneously. Single- stream variants are easier entry points; the dual version is the demanding one.
- Sensible ceilings. Spatial N-back plateaus around 4–5 back for nearly everyone; auditory streams cap lower. Apps that let N grow forever are padding, not progressing.
N-back in Avoken
Avoken ships three N-back variants — an identity stream, a spatial stream on a grid, and an auditory stream — each with planted match quotas and lures at upper levels. Difficulty steps are calibrated to the published paradigm rather than tuned to keep you tapping, and each workout cites its sources in-app.
Avoken is a training app, not a treatment — we describe the task, and we don't claim it transfers anywhere else. If you like the paradigm, it's a careful implementation.