A meta-analysis published in Aggression and Violent Behavior in 2024 reports that omega-3 supplementation produced a modest short-term reduction in aggressive behaviour across 29 randomised controlled trials, covering 3,918 participants and 35 independent samples from 19 laboratories. The lead author, University of Pennsylvania neurocriminologist Adrian Raine, has summarised the headline finding as roughly a 30 percent reduction in aggression, holding up across age, gender, diagnosis, treatment duration, and dosage. Most participants in the included trials were not violent offenders — the finding covers community and clinical samples across a wide age range.
The finding is worth taking seriously, but it should not be read as the final word. This is a meta-analysis of short-term trials averaging sixteen weeks, with the lead author also responsible for five of the included studies. The effect is described by the authors themselves as modest.
What the paper actually measured
The dataset covers trials conducted between 1996 and 2024, in a range of populations: children aged sixteen and under, adults in middle age, clinical samples, and community samples. Treatment groups received omega-3 fatty acids, usually as fish oil, and were compared with control groups over the trial period. Aggression was measured through a mix of self-report instruments, parent and observer reports, and clinical scales, depending on the original study.
The reported reduction held when the authors averaged effect sizes by study, by independent sample, and by laboratory. That triple check matters. It is a reasonable defence against the result being driven by one large outlier trial or by a single research group.
The detail that has drawn the most attention is the split between reactive and proactive aggression. Reactive aggression is the heated kind, the kind that follows provocation. Proactive aggression is the cooler kind, planned in advance for instrumental ends. Earlier work had left it unclear whether omega-3 affected one type, the other, or both. In this meta-analysis, the authors report a reduction in both.
That is the cleanest novel contribution of the paper. It does not prove the mechanism. It says the statistical signal shows up on both sides of a distinction that aggression researchers have spent decades trying to keep separate.
What the paper does not show
Several things, and worth naming plainly.
It does not show long-term effects. Only one of the nineteen contributing laboratories followed participants after supplementation ended, according to the paper’s own account of the included trials. The analysis is bounded by the duration of treatment, on average sixteen weeks. Whether the effect persists, decays, or reverses once the capsules stop is, on this evidence, unknown. The authors say so in the paper.
It does not establish a mechanism. The discussion section gestures toward omega-3’s role in reducing inflammation and supporting prefrontal function. These are plausible candidate pathways, supported by separate bodies of research, but the meta-analysis itself was not designed to test them. The behavioural result is independent of any settled story about why.
It does not control for the unusual position of the lead author. Raine has been publishing on omega-3 and aggression for years and contributed five of the twenty-nine trials. This is openly stated in the paper and the Penn release, and the meta-analysis was conducted with co-author Lia Brodrick of the Perelman School of Medicine. It is not a hidden conflict. It is, however, a reason to want the result replicated by groups with no prior commitment to the hypothesis.
And it does not address dose-response with much resolution. The trials varied widely in how much omega-3 was given, in what ratio of EPA to DHA, and to what kind of participant. The meta-analytic finding rolls all of that together. It tells you the average direction of the effect, not the dose at which to expect it.
The criminal-justice framing, and its limits
Raine is a neurocriminologist, and the framing of the Penn release reaches in that direction, talking about the community, the clinic, and the criminal justice system. The paper itself recommends that omega-3 supplementation be considered as an adjunct to existing psychological interventions such as CBT and pharmacological interventions such as risperidone, with caregivers informed of the potential benefit.
This is a careful recommendation. The word adjunct is doing real work. The authors are not suggesting that omega-3 replaces therapy or medication for people whose aggression is severe enough to warrant either. They are suggesting it sit alongside.
The wider cultural reading is at risk of being looser. The temptation, when a headline number lands at 30 percent, is to picture it as a uniform reduction in fights, violence, and antisocial behaviour across a population. The paper does not support that picture. The effect is averaged across many measures of aggression in many populations, most of whom were not violent offenders. What the trials show, in aggregate, is that scores on aggression instruments come down in supplemented groups relative to controls, over weeks, by a modest amount.
That is interesting. It is not the same as a public-safety intervention with a known effect at the population level.
Why this lands differently from most diet-and-behaviour stories
Nutrition research on behaviour has a poor record, generally. Single trials produce headline effects that fail to replicate. Cross-sectional studies confuse cause and effect. Industry-funded work on specific supplements tends to find what the industry hoped it would find. Readers have reasonable cause to be tired.
The Raine meta-analysis is closer to the better end of this literature for a few reasons. It pools randomised trials rather than observational data, which addresses the most basic causality problem. It triangulates the effect three different ways. It reports a modest effect rather than a sensational one. It names its limitations. And it is published in a specialist journal in the field where the outcome variable — aggression — has been the subject of decades of measurement work.
None of that makes it definitive. It does mean the finding sits a step or two further along the evidence ladder than the typical wellness-supplement story.
What this is not advice to do
The paper is not a recommendation to medicate ordinary irritability with fish oil. It is not a substitute for the things that already work for aggression in clinical or forensic settings: behavioural therapy, in some cases medication, structured environments, and the long, unglamorous work of skill-building. Raine himself is clear that omega-3 is not a magic bullet.
It is also not a reason to read alcohol, sleep, or exercise out of the picture. Nothing in the paper argues against any of the established behavioural inputs that affect mood and self-regulation. The trials add one input to the list, with a modest average effect, in the short term, on a measurable outcome.
For people thinking about their own diet, the ordinary public-health message about fish intake has not changed. As Raine put it in the Penn Today release: “At the very least, parents seeking treatment for an aggressive child should know that in addition to any other treatment that their child receives, an extra portion or two of fish each week could also help.” The paper offers some support for taking that message seriously. It does not turn fish oil capsules into a treatment.
What we will be watching
The authors flag the next steps clearly. Long-term follow-up after supplementation ends. Brain-imaging work on whether omega-3 changes prefrontal functioning in the way the mechanistic story would predict. Genetic moderators. And whether self-report and observer-report measures of aggression converge or diverge under treatment.
If those studies arrive and broadly hold up, the case Raine is making becomes harder to dismiss. If they do not, this meta-analysis will sit in the literature as a careful pooled estimate of a real but small short-term effect, useful mostly as an adjunct in clinical settings where aggression is already being addressed by other means.
The paper itself is honest about which of those futures it is pointing toward. The coverage around it will not always be.
This article is a reading of the published research and surrounding coverage, not medical advice. For readers whose own aggression, or a family member’s, has become a source of real concern, a GP or qualified mental health professional remains a more useful first call than any single study or supplement.
