Working memory

Holding a phone number briefly is storage. Reordering it, combining it with another number or replacing its first digit adds control. Working memory is the temporary workspace that supports those operations. Long-term knowledge supplies familiar words, patterns and strategies, so the same sequence can be easier for an expert than a beginner. Capacity depends on what counts as a meaningful chunk. 2

In a simplified brain-system account, frontal and parietal networks help maintain goals and direct attention, while sensory and language systems help represent the particular material. There is no single working-memory box, and a browser score cannot tell us which brain area was more active. Executive control and working memory overlap; they are not interchangeable labels for every difficult task. 3

Helibrain lets you approach this distinction from several angles. Number memory asks for a sequence back; n-back asks whether the current item matches an earlier one; visual and sequence games use locations. Those differences matter more than which game produces the largest-looking score. A strong result in one format can coexist with a modest result in another without either being wrong. 4

These are related measurements, not interchangeable units. Complex-span and n-back tasks share only part of their individual-difference variance. Calling a high n-back level a digit span, or converting one into an IQ, skips the validation that would be needed. 4

Our visual number-memory game is not the auditory protocol in Woods and colleagues’ digit-span study. Their work shows that maximum span, total correct and an adaptive mean-span measure give different answers even within one task family. Keep the rule and scoring method attached to a result. 5

The table below reproduces scaled digit-span scores, not numbers of digits, from Ye and colleagues’ Chinese normative sample. The scale was formed across all 723 participants, with mean 10 and standard deviation 3; these age-group means are not separately age-normalized scores. Participants were Mandarin-reading adults aged 18–59. 6

Source: the article’s age-group table, panel A. Education and the particular digit-span procedure matter. None of these cells is a percentile lookup for Helibrain. A separate large online study found working-memory task peaks around age 30, illustrating that the chosen task and sample alter the curve. These are group comparisons, not a forecast of an individual’s future. 7

Learning a task can make you better at that task. A harder question is whether the improvement reaches unfamiliar tasks that use similar processes. The n-back training literature shows a gradient: performance changes depend on how similar the outcome is to the training. It does not justify treating a higher game level as a larger general mental capacity. 11

Far transfer means improvement in a more distant outcome, such as intelligence, academic achievement or everyday functioning. A major working-memory training meta-analysis did not find convincing broad far-transfer benefits. Helibrain’s games therefore offer practice and a way to observe task performance; we do not present them as a way to raise intelligence. 13

A practical distinction is between changing your capacity and changing the demand. A note, diagram or written intermediate calculation can reduce what you need to keep active. That is a sensible use of an external aid, not evidence that the aid changed the size of working memory. Our concept-map and study tools support organizing a task; their usefulness should be judged on that task.

Start with a storage task and an updating task, then read their separate explanations. Record the input mode, whether you rehearsed aloud or grouped items, and whether the task was familiar. A “best ever” score after many attempts answers a different question from a typical first-session score.

The cards below distinguish direct task practice from tools that help you organize material. They do not share a combined working-memory percentile. If you want a personal comparison, keep conditions similar and look for a pattern across sessions rather than a single impressive run.