Research and news

In 235 children aged 6 to 13, eight weeks of response-inhibition training still helped a closely related control score a year later. School achievement, fluid reasoning, mental health, and the brain measures did not change.

This review of trials in healthy older adults found little evidence that computerised cognitive training lasting 12 weeks or more improves cognition. The authors rated that evidence as low quality.

Their meta-analyses of cognitive training, music, board games, and video games found little lasting gain in general cognitive ability once study quality and statistical artefacts were taken into account.

Across 42 data sets and more than 600,000 people, an extra year of schooling lined up with about 1 to 5 IQ points. The association showed up on several kinds of tests and was still present later in life.

Twin studies put about half of the differences in adult intelligence down to genetic differences. Genome-wide scores available at the time captured about a fifth of that genetic component.

A review of studies cited by brain-training companies found reliable gains on the practiced tasks, weaker evidence for similar tasks, and little evidence that the practice helps everyday life.

Lumos Labs agreed to pay $2 million and to stop claiming that Lumosity games improve school, work, or athletics, or that they delay dementia, unless it has competent scientific evidence.

Dozens of cognitive scientists said advertised brain games did not have compelling evidence for improving everyday thinking or preventing cognitive decline. They asked for careful trials instead of broad claims.

ICAR is a public-domain ability test. This paper describes the first item types and how the scores line up with other ability measures. The 16-item form on this site comes from that work.

Ten years after memory, reasoning, or speed training, each trained group reported less difficulty with daily tasks than the control group. Reasoning and speed still showed an effect on the trained skill. Memory training did not.

Twenty sessions of adaptive dual n-back made people better at dual n-back. Compared with an active placebo, fluid intelligence and the other ability tests did not improve.

Working-memory programs raised working-memory scores in the short term. The review found no convincing carry-over to reasoning, attention, word reading, or arithmetic.

This review treats intelligence-test scores as useful, traces how both genes and environment contribute, and argues that scores can change more across a life and a generation than early theories allowed.

11,430 people trained online for six weeks. They improved on every task they practiced. Those gains did not appear on untrained tests, including tests that were closely related.

The review treats a century of testing as support for a general factor of intelligence. Imaging studies link higher scores with larger brains and with networks spread across frontal and parietal cortex, not one small region.

Adults who practiced a demanding working-memory task later scored higher on matrix reasoning, a test they had not practiced. People who trained for more days gained more. Later trials have not all found the same transfer.

Young chimpanzees remembered the positions of numerals shown for a fraction of a second better than university students on the same apparatus. The Chimp Test on this site is that kind of task, written for people.

In 2,832 adults aged 65 to 94, ten sessions of memory, reasoning, or speed training improved the skill that was trained, and the gain was still measurable two years later. Everyday functioning had not changed at two years.

Across 85 years of hiring studies, general mental ability strongly predicted job and training performance. Pairing it with a work sample or a structured interview predicted better than either piece alone.

General intelligence predicts performance best when the task is complex. The paper uses adult literacy data to show the same pattern in ordinary tasks such as forms, schedules, maps, and news articles.

Scores on intelligence tests rose by a large amount across the twentieth century in 14 countries. The rise was bigger on some tests than others, which is now called the Flynn effect.

A target letter is identified more slowly when the letters beside it call for the opposite response. That flanker effect is the task behind Middle Arrow.

Cattell separated fluid ability, reasoning on new problems, from crystallized ability, skill and knowledge already learned. The two are related, and they do not move in lockstep across the life span.

People had to answer about a signal that had appeared one or more steps earlier, while new signals kept arriving. Older adults held less of that recent stream. The n-back task used on this site comes from this running-memory idea.

Naming the ink of a color word is slower when the word names a different color. That interference is the Stroop effect, and it is the task in the Stroop Test here.

Scores on different mental tests rise and fall together. Spearman argued that the shared part is one general factor, which he called g, plus a factor special to each test.