Human neuroscience is at a crossroads where both replicability and generalizability of neurocognitive research remain highly challenging due to potentially unprecedented complexity of brain-behavior associations and thus hamper their translations into healthcare practices. An important aspect of such complexity is individual differences while population neuroscience provides a promising framework for addressing this complexity with an interdisciplinary paradigm by integrating neuroscience, genetics and epidemiology. A fairly attractive research direction in developmental population neuroscience is to uncover the patterns, origins and neural mechanisms of these individual differences in the human brain

and mind. The complexity of understanding these individual differences in this new direction presents great challenges for current research methods. To fill this methodological gap, we argue that there is an urgent need to develop a methodological framework for modelling individual differences at both the population and individual levels with quantitative general laws and the characteristics of individual differentiation. Normative modelling provides an effective methodological choice for this purpose to bridge general laws and individual translations. This method can be used to generate statistical models at the population level by leveraging distributional regression approaches on representative samples and derives deviations at individual level from the population reference models as individual differentiation quantities. This methodological framework can provide a novel tool for developmental population neuroscience research through which common patterns and the biological underpinnings of individual differences in the brain and mind can be elucidated from a lifespan perspective. This framework offers reference resources that can be used to not only reveal possible etiological mechanisms of various brain-mind disorders but also implement their precise clinical practices such as the pediatric growth charts for references and standards. According to the vast potential of normative modelling methods, in this review, we encourage the immediate action of optimizing sampling strategies, standardizing data collection, and improving statistical modelling for deep investigations to accelerate their translational applications. In recent years, China has accumulated rich resources of in-vivo neuroimaging data, paralleling to the international brain imaging data sharing and open brain science research. Looking beyond, various large-scale brain-mind cohorts will be continuously initiated to cover different regions and different stages of lifespan through the program of scientific and technological innovation 2030-The major project of the Brain Science and Brain-Inspired Intelligence Technology. This will offer China the world's largest living brain-mind resources for promoting the research, application and transformation of normative modeling methods in the development of population neuroscience. Such advances will lead to construct normative brain-mind association models for the lifespan development of Chinese population, assist the early diagnosis and precise treatment as well as reform the clinical practice. It is expected to ultimately provide scientific support for parsing the etiology of the brain-mind diseases, and optimizing the national brain health and its related policy designation.

摘要揭示脑与心智个体差异的来源、规律及其科学机制是发展人口神经科学领域颇具吸引力的研究方向.个体差异的复杂性与当前研究方法的局限性是该领域方向的挑战所在，亚须发展在群体和个体层面建模个体差异的方法学框架，来量化脑智群体规律及其个体分化特征，实现一般科学规律与个体应用转化的科学衔接.规范化建模为此提供了一种方法学选择:基于代表性样本，使用分布式回归等方法，建立群体规律的统计模型，并在此基础上推演个体偏离规范化群体参照的程度，从而量化个体分化指标.这一方法学框架为发展人口神经科学提供了新工具，从人口生命周期的视角揭示了个体脑智差异的一般规律及其生物学基础，为解析脑智疾患的病因学机制及精准化临床实践提供了参考.鉴于规范化建模方法的巨大潜力，未来亚须在采样策略优化、数据收集规范以及建模方法改进等方面开展深入研究，从而加速其应用转化.