What a Child’s Blink May Reveal — and What It Cannot

Sort the red rabbit and blue boat by color. Then change the rule and sort them by shape. To an adult, the instruction sounds trivial. To a three-year-old, it can expose one of the most striking transitions in early cognitive development: a child may understand the new rule and still continue applying the old one.

A Japanese research team has now connected that familiar preschool challenge to an unexpectedly ordinary physiological signal — spontaneous blinking.

Ryuta Kuwamizu of the University of Tsukuba and Yusuke Moriguchi of Kyoto University, with Nozomi Yamamoto and Kota Otani, studied rule switching, prefrontal activity and spontaneous eye-blink rate in 113 young children. Their peer-reviewed paper was published September 15 in Communications Psychology. ^{[1] [2] [3]}

Children who blinked more frequently tended to perform better on the rule-switching measure. Blink rate was also associated not with broad prefrontal activation, but specifically with activity in the right dorsolateral prefrontal cortex, or rDLPFC — a region repeatedly implicated in cognitive control. Older children showed a stronger rightward and dorsal concentration of prefrontal activation. ^[1]

Executive function: the mind’s control system

Executive function is an umbrella term for cognitive processes that allow people to regulate thought and behavior in pursuit of a goal. It commonly includes inhibitory control, working memory and cognitive flexibility — the ability to change rules, strategies or perspectives when circumstances change.

For a preschool child, these capacities are not abstract laboratory concepts. They are involved in stopping one activity when a teacher gives a new instruction, remembering a two-step direction, taking turns and adapting when the rules of a game change.

Executive function develops rapidly in early childhood, and previous longitudinal research has linked individual differences in these skills with later academic, behavioral and health outcomes. Those associations do not mean a preschool task determines a child’s future; development is shaped by many biological, educational, family and social factors. The new study addresses only one part of that much larger system. ^{[1] [2]}

The card-sorting task that made preschool inflexibility famous

The researchers used the Dimensional Change Card Sort, or DCCS, one of developmental psychology’s best-known tests of cognitive flexibility. Cards vary on two dimensions — often color and shape. Children sort by one rule, then are told to use the other.

The task emerged from developmental work in the 1990s and was standardized in a widely used 2006 protocol by Philip Zelazo. In the classic pattern, many three-year-olds perseverate on the original rule after the switch, while by about age five most typically developing children can switch successfully. A later meta-analysis documented the task’s extensive use across developmental studies and the many theories proposed to explain why the age transition occurs. ^{[4] [8]}

The 2026 experiment used a more demanding multistage form. Each of three sessions contained a pre-switch phase, a post-switch phase and a mixed phase in which color and shape rules alternated from trial to trial. Each task block lasted 25 seconds, separated by 15 seconds of rest. ^[1]

Who were the children?

The final sample consisted of 55 girls and 58 boys recruited from two preschools in Osaka and Kyoto. Mean age was 59 months; the full range was 35 to 80 months. The preschools reported no developmental abnormalities among participants. Parents provided informed consent, and the study received ethics approval at Kyoto University. ^[1]

An additional 10 children did not yield usable data: six declined to start the experiment, three refused to wear the fNIRS cap and one could not complete the task because of excessive movement. That small detail illustrates a recurring issue in developmental neuroscience — data necessarily come from children who can and will tolerate the measurement procedure. ^[1]

Watching the brain without putting a preschooler in an MRI scanner

The study used functional near-infrared spectroscopy, or fNIRS. The method shines near-infrared light through the scalp and tracks task-related changes in oxygenated and deoxygenated hemoglobin. Those hemodynamic changes provide an indirect measure associated with neural activity.

For young children, fNIRS has a practical advantage over MRI: it does not require lying motionless inside a large, noisy magnet. Children can sit at a table and perform a task while wearing a lightweight array of emitters and detectors.

But fNIRS has limitations. It samples cortical surface regions rather than the whole brain, has lower spatial precision than MRI, and its signal can contain non-neural contributions from scalp blood flow and other systemic physiology. The authors explicitly discuss those constraints and used frontal placement, quality control and signal-separation methods to reduce contamination. ^[1]

The 16-channel arrangement covered bilateral dorsolateral and rostrolateral prefrontal cortex. The researchers analyzed changes in oxygenated hemoglobin across task phase, hemisphere and cortical subregion. ^[1]

A Japanese research line that reaches back to 2009

The study extends a long-running program in Japanese developmental neuroscience. In 2009, Moriguchi and Kazuo Hiraki used near-infrared spectroscopy while three-year-olds, five-year-olds and adults performed a card-switching task. They reported that successful switching in young children was associated with prefrontal activation; three-year-olds often struggled after the rule changed, while five-year-olds and adults performed much more successfully. ^[5]

Later work followed prefrontal development longitudinally. In 2018, a Kyoto University group led by Moriguchi reported associations among variation in the dopamine-regulating COMT gene, executive-function performance and lateral prefrontal activation in children aged three to six. ^[6]

The new blink study adds a different question to that history: is there a simple, observable physiological signal that covaries with the same developing cognitive and prefrontal system?

How do you measure a blink?

The children’s faces were recorded by a front-facing video camera at 30 frames per second, positioned about 100 to 150 centimeters away. Spontaneous eye-blink rate was calculated across the roughly six-minute DCCS session.

Two trained raters independently counted blinks for every video, and the researchers averaged their counts. Agreement was high: both the simple correlation and intraclass correlation coefficient were 0.93. A third rater rescored a random 20% of recordings and again showed very high agreement. ^[1]

The study focused on spontaneous blinking — the ordinary blinks that occur without a deliberate instruction or an external trigger. Earlier literature cited by the paper reports very low spontaneous blink rates in newborns, roughly 0–5 per minute, followed by a several-fold increase through early childhood toward adult rates around 20–30 per minute. In the new sample, blink rates were uniformly low with little variation below age four, while older children showed much wider individual differences and some already exceeded 20 blinks per minute. ^[1]

Finding one: older children switched better

Age was strongly associated with switching accuracy: Spearman’s rho was 0.68. That result is consistent with decades of DCCS research showing rapid improvement between the preschool years. ^[1]

Blink rate also increased with age, but more modestly: rho was 0.29. The critical test was whether blinking related to performance over and above age.

It did, although the association was moderate rather than large. Blink rate and switching accuracy correlated at rho=0.35. After statistically controlling for age, the relationship remained at partial rho=0.22. ^[1]

Finding two: prefrontal activity became more selective with age

During the post-switch and mixed phases, prefrontal activation was greater than during the initial-rule phase. Across the sample, activity was stronger in the right hemisphere than the left and stronger in dorsolateral than rostrolateral prefrontal cortex. ^[1]

Age added another pattern. Older age was associated with less activity in left and rostral regions but more activity in right and dorsal regions. The authors interpret this as emerging functional differentiation: the developing brain may shift from broader recruitment toward more selective use of regions well suited to the task. ^[1]

That idea fits a broader developmental-neuroscience literature describing diffuse-to-focal or increasingly specialized patterns of cortical recruitment across childhood.

Finding three: blinking converged on the right DLPFC

The researchers then asked whether blink rate and switching accuracy related to activity in each of four measured prefrontal regions: right and left DLPFC and right and left rostrolateral PFC.

After correction for multiple comparisons, only the right DLPFC showed a significant positive correlation with both measures. Its activity correlated with blink rate at rho=0.27 and with switching accuracy at rho=0.32. The other regions did not meet the corrected significance threshold. ^[1]

Positive relationships remained when age was statistically controlled. The authors also report robustness checks adjusting for sex and preschool site and excluding influential outliers. ^[1]

What this does not establish is causal direction. The rDLPFC does not necessarily “cause” more blinking, and blinking does not train the rDLPFC. An exploratory mediation analysis was consistent with a shared developmental account, but the authors explicitly call for further mechanistic work. ^[1]

The dopamine story is tempting — and incomplete

Spontaneous eye-blink rate has a long history as a proposed indirect correlate of central dopaminergic function. Animal, pharmacological and clinical studies have found relationships between blinking and dopamine systems, and a 2016 review summarized evidence linking blink rate with dopamine-related cognition, including cognitive flexibility. ^[7]

That history makes dopamine an obvious candidate mechanism. Prefrontal and striatal dopamine systems also play important roles in task switching and updating.

But the new paper is unusually explicit that blink rate is not a specific dopamine meter. Blink generation involves brainstem and subcortical structures and is likely influenced by several neuromodulators involved in arousal, including dopamine, acetylcholine and noradrenaline. Studies in healthy adults have also failed to reproduce a simple positive relationship between blink rate and some direct measures of dopamine synthesis or receptor availability. ^[1]

Could parents or teachers use this as a developmental test?

The appeal is obvious. A blink can be recorded with an ordinary camera rather than expensive neuroimaging equipment. The authors argue that, if replicated and developed, blink-rate measurement could support large-scale research in homes, preschools and schools and could eventually be paired with automated video analysis. ^[1]

But there is a large gap between a promising research marker and a usable test of an individual child.

fNIRS does not see the dopamine system directly

There is another important boundary. The experiment measured lateral prefrontal cortex at the surface. It did not directly record activity in deeper structures such as the striatum or midbrain, which are central to many dopamine theories of executive function. ^[1]

Nor does a change in oxygenated hemoglobin equal a direct recording of neuronal firing. fNIRS is a hemodynamic measure. The authors note that extracerebral and systemic physiology can contribute to the signal, even though their processing was designed to reduce those effects. ^[1]

Head circumference was not measured individually, creating another possible source of developmental variation in channel placement. The authors estimate that average Japanese head growth from age three to six would shift central channels only by a few millimeters and argue that their regional analysis should absorb most of that difference. ^[1]

Why blinking is still worth taking seriously

The practical challenge of early-childhood neuroscience is simple: many of the best brain-measurement techniques ask young children to do things young children are least suited to do — remain still, tolerate sensors and follow long protocols.

Blinking requires almost nothing from the participant. If its relationship with prefrontal functional development proves reliable across tasks, cultures and longitudinal samples, its very ordinariness could become its scientific strength.

The present study does not turn a webcam into a brain scanner. It suggests something subtler: a visible behavior that usually passes unnoticed may covary with a hidden developmental process that is otherwise difficult and expensive to measure.

The next test is time

The strongest next experiment would follow the same children over several years. Does a child’s spontaneous blink rate rise in parallel with the increasing specialization of prefrontal activity? Does that trajectory predict improvement across several executive-function tasks, not just card sorting?

Researchers will also need to test broader populations and account for factors that can influence blinking or arousal — fatigue, sleep, attention, visual conditions, eye-surface irritation, screen exposure and moment-to-moment internal state. The paper itself notes that its average blink-rate measure does not capture the fine timing of individual blinks, which may carry different information. ^[1]

The study therefore leaves us with a useful distinction. One child blinking more often than another tells us very little by itself. But across 113 children, when blinking, behavior and prefrontal hemodynamics are measured together, a developmental pattern appears.

That pattern is not a diagnosis. It is a new window — small, ordinary and opening several times a minute.