The "Reading the Mind in the Eyes" Test in Autism-Spectrum Disorders Comparison with Healthy Controls: A Systematic Review and Meta-analysis.

The team pooled 18 studies that used the "Reading the Mind in the Eyes" Test.

They compared people with autism to neurotypical peers.

They also looked at which IQ scores predicted success for each group.

Neurotypical people scored higher on the test.

For them, verbal IQ predicted better scores.

For people with autism, only performance IQ mattered.

This hints at different brain routes to the same answer.

Ivancic et al. (1996) saw a similar split. Adults with intellectual disability read faces in their own way, not just more poorly.

Kemner et al. (2008) looked positive at first glance: adults with PDD found hidden objects faster than controls. That seems to clash with the lower RMET scores. The twist is task type. Visual search rewards detail focus; RMET rewards social inference. Different strengths, not a contradiction.

Dickinson et al. (2014) linked sharper line-angle judgment to higher autistic traits in neurotypical adults. Together these papers build one story: autism can bring visual strengths, yet social-eye cues remain hard.

When you give the RMET, pair it with performance IQ, not just language tests. A low score may reflect strategy, not lack of empathy. Use this insight to pick teaching tools that lean on visual logic rather than verbal hints.

The Reading the Mind in the Eyes Test (RMET), developed by Simon Baron-Cohen and colleagues, shows photographs cropped to just the eye region. For each photo the person chooses one of four mental-state words that best describes what the person in the picture is thinking or feeling. The revised adult version has 36 items, and a separate children's version exists.

The RMET is meant to measure the ability to infer emotional and mental states from subtle facial cues, sometimes called affective theory of mind or advanced emotion recognition. It taps recognition of complex states like skeptical, cautious, or contemplative, not just basic emotions such as happy or afraid.

It is a research and screening instrument, not a diagnostic test. A score cannot confirm or rule out autism on its own, because performance also depends on vocabulary, attention, and cultural familiarity with the stimuli.

On average, autistic adults score lower on the RMET than non-autistic controls, and this meta-analysis of 18 studies confirmed that pattern. But the more useful finding was about how RMET performance related to intelligence. In controls, RMET scores were tied to full-scale and verbal IQ and to age. In autistic participants, RMET scores were negatively correlated with performance IQ.

That difference suggests autistic and non-autistic people may approach the task in different ways. Control participants appear to lean on verbal ability, while the pattern in autistic participants implies their performance depends less on verbal skills. In other words, the test may not measure the same underlying process across groups.

This matters for interpretation. A given RMET score does not carry the same meaning for an autistic adult as for a non-autistic adult, so scores should be read alongside a person's IQ profile rather than compared to a single norm.

The adult RMET is scored out of 36, and most non-autistic adults land somewhere in the mid-20s. There is no clinical cutoff that marks a person as autistic, and the test is not a substitute for a full diagnostic evaluation.

Because scores are influenced by IQ, vocabulary, age, and the cultural background of the photographed faces, a single number should be treated as one data point, not a verdict. Use the RMET as a research measure or a conversation starter, and interpret any individual score with the person's broader profile in mind.