Aggression affects nearly all mental disorders and constitutes a burden for society in general and the mental health sector in particular. An incomplete understanding of its underlying neurobiological mechanisms limits treatment efficacy and hinders the development of precision medicine approaches. Here we review recent efforts employing neuroimaging, hormonal and (epi)genetic data to develop biomarkers for aggression. Our goal is to discuss limitations of these previous studies and propose strategies which may help overcome these limitations in order to achieve both: transdiagnostic classification and prediction of individual treatment outcomes. In order to develop clinically relevant biomarkers for aggression, candidate biomarkers have to be combined and their reliability and validity has to be tested rigorously in large patient cohorts. Besides generating a better mechanistic understanding, we should strive to create multimodal biomarkers that medical professionals can utilize in clinical practice to guide and improve the management and clinical outcome of pathological aggression.

攻击行为几乎影响所有精神障碍，对整个社会尤其是精神卫生部门构成负担。对其潜在神经生物学机制的不完全了解限制了治疗效果并阻碍了精准医学方法的发展。在这里，我们回顾了最近使用神经影像学、激素和（表观）遗传数据开发攻击性生物标志物的努力。我们的目标是讨论这些先前研究的局限性，并提出可能有助于克服这些局限性的策略，以实现两者：跨诊断分类和个体治疗结果的预测。为了开发临床相关的攻击性生物标志物，必须结合候选生物标志物，并且必须在大型患者队列中严格测试其可靠性和有效性。