Lemon balm: the calm studies, and what they actually measured
Lemon balm is a lemon-scented mint that has been grown in kitchen gardens across Europe and the Mediterranean for centuries. The modern research on it is small, mostly single-dose, and did something most herb studies do not: it measured what happened within hours rather than weeks.
The single-dose studies
A crossover study gave twenty healthy young adults single doses of 300, 600 and 900 mg of a standardised extract. The 600 mg dose produced a sustained improvement in accuracy of attention. The 300 mg dose raised self-rated calmness at the earliest time points. The 900 mg dose reduced alertness at every time point, and some doses reduced secondary and working memory scores (Kennedy et al., Pharmacology Biochemistry and Behavior, 2002, PMID 12062586).
That last part is worth pausing on. More was not better. The highest dose was the one that dulled people.
A second crossover study put eighteen volunteers through a twenty-minute laboratory stressor after a single dose. At 600 mg, the negative mood effects of the stressor were reduced and self-rated calmness increased, with alertness reduced. At 300 mg, speed of mathematical processing increased with no loss of accuracy (Kennedy et al., Psychosomatic Medicine, 2004, PMID 15272110).
Two different doses, two different profiles: the lower one sharper, the higher one calmer and slower.
A longer study, with a weaker design
An open-label study followed twenty adults with mild to moderate anxiety and disturbed sleep for fifteen days on a standardised extract. Anxiety manifestations reduced by 18%, associated symptoms by 15%, and insomnia by 42%, with nineteen of the twenty classed as responders (Cases et al., Mediterranean Journal of Nutrition and Metabolism, 2011, PMID 22207903).
Those numbers look impressive and the design does not support them carrying much weight. There was no placebo group, so there is nothing to compare the improvement against, and people who enrol in a study and take something daily for a fortnight tend to report feeling better.
The studies in children, and why they prove less than they appear to
Two studies did involve children, and both are commonly cited without their caveats. They deserve the caveats.
An open multicentre study of 918 children under twelve with restlessness and trouble falling asleep reported that difficulty sleeping improved in 80.9% of those affected and restlessness improved clearly in 70.4%, with tolerability rated good or very good in 96.7% (Müller and Klement, Phytomedicine, 2006, PMID 16487692). An observational study of 169 primary school children with hyperactivity and concentration difficulties not meeting the threshold for a diagnosis reported large falls in the proportion with strong symptoms over seven weeks (Gromball et al., Phytomedicine, 2014, PMID 24837472).
Three things about both. Neither had a control group, so there is no way to separate the preparation from time, attention and expectation. Both used a fixed combination of valerian and lemon balm, not lemon balm alone, so no result can be attributed to lemon balm. And both were authored by people associated with the manufacturer.
They are worth knowing about. They are not evidence that lemon balm on its own does anything in children, and anyone citing them as if they were is not reading them.
The fair summary
Lemon balm has small, well-designed adult studies showing changes in self-rated calmness and attention within hours of a single dose, with a dose response that is not simply more-is-better. In children it has two large uncontrolled studies of a valerian combination, which is a much weaker thing than it sounds.