Anxiety and Self-Control in Children Reinforce Each Other

What the review found

Whether a child’s anxiety produces their difficulty with planning and self-control, or whether the difficulty produces the anxiety, is a live dispute with long-term evidence on both sides. Vulnerability models place the cognitive difficulty first and treat worry as what grows up around it. Scar models reverse the order and treat worry as the thing that wears the cognitive machinery down over years.

Nur Hani Zainal, at Harvard Medical School, and Michelle G. Newman, at Pennsylvania State University, set out that dispute in a 2022 review for Current Psychiatry Reports. Their paper is narrative rather than statistical, with no systematic search and no combined effect estimate, so every result it reports is described in words rather than numbers. The evidence runs from infancy through adults over 100, most of the long-term work in adults.

Executive functioning here covers the control processes: stopping an automatic response, switching between rules or tasks, holding information in mind, planning. Measurement at a single point in time says nothing about which came first, so the review rests on studies that follow the same people over years.

On the vulnerability side, parent-reported behavioural inhibition in toddlers predicted socially anxious behaviour in middle childhood, with reduced set-shifting mediating between the two. Executive dysfunction and slow processing speed in children aged nine to twelve predicted behavioural and emotional problems two years later, and in community adults the same class of difficulty predicted generalized anxiety disorder nine years later, with baseline scores controlled.

The scar evidence is as specific. Anxiety in early adolescence predicted working memory problems in late adolescence and not the reverse, and a rise in pathological worry within one person over nine years predicted a subsequent nine-year reduction in inhibition and set-shifting from midlife into older adulthood.

A pattern that holds when comparing different people does not necessarily hold when tracking one person over time, and the within-person results are the stronger evidence. In children aged six to fourteen, anxiety severity predicted declines in parent-rated executive function at both levels, and executive function predicted anxiety in return. Kindergarteners’ rising worries about sleep predicted falling planning, inhibition and shifting, and the reverse was also true.

The same reciprocity appeared in adolescents, and in adults aged forty to eighty-four growing worry coincided with decreases in verbal working memory and processing speed, again in both directions. In adults aged seventy to one hundred and ten, a separate analysis found no such within-person relation at all.

The overgeneralized control model runs against that symmetry: higher inhibition in toddlerhood predicted more anxiety ten years later, most strongly in children whose proactive control was weaker, and enhanced cognitive control accompanied greater anxiety in some youths.

What it means for you and your child

If a report describes a child as both anxious and effortful in planning or in holding instructions, the order of events a parent reconstructs may not be recoverable. The review presents its five accounts as concurrently supported rather than as candidates for a single winner, and describes the evidence of temporal precedence as sufficient only for weak causal inferences.

Improvement on one side does not guarantee movement on the other. Forty sessions of cognitive behavioural therapy did not alter cognitive function in patients with major depressive disorder, though the mismatch between how those patients rated their thinking and how they performed on tests had disappeared by the end of treatment. Those were depressed adults rather than anxious children, and the review reports no equivalent test in children.

Inflammatory markers in the blood, including interleukin-6 and C-reactive protein, mediated the relation between excessive worry and reduced executive function two to eighteen years later in two separate samples, and did not do so for panic symptoms. Elevated daily cortisol, the more obvious candidate, mediated nothing. This is a lead rather than an explanation, resting on two samples and one failed alternative.

What has randomised support is unglamorous. Trial evidence is accumulating that regular physical activity, twenty to one hundred and twenty minutes, one to five times a week, improves executive function. Working memory training built on neutral rather than distressing material reduced repetitive negative thinking in young healthy adults and habitual worriers, and a meta-analysis found promise for mindfulness-based interventions, primarily in healthy controls.

Cross-sectional studies saturate the field, few published vulnerability studies report null results, and the long-term work needed to test mechanisms, with three or more measurement points, remains scarce. Nothing in the review establishes that either condition produces the other in a given child, and its authors do not claim otherwise.

The intervals in these studies are years rather than weeks: two years, nine years, a decade. Change measured on that scale is invisible from inside a single school term, and a parent watching for the moment when one condition releases its hold on the other is watching for an event this literature has never recorded. What it does record is slow movement in both directions at once, which means there is no first thing to fix, and no sequence you are behind on.