Mareva Silvana, Holmes Joni
MRC Cognition & Brain Sciences Unit, University of Cambridge, Cambridge, UK; Psychology Department, Faculty of Health and Life Sciences, University of Exeter, UK.
MRC Cognition & Brain Sciences Unit, University of Cambridge, Cambridge, UK.
Cortex. 2024 Mar;172:204-221. doi: 10.1016/j.cortex.2023.11.021. Epub 2024 Jan 23.
Executive function, an umbrella term used to describe the goal-directed regulation of thoughts, actions, and emotions, is an important dimension implicated in neurodiversity and established malleable predictor of multiple adult outcomes. Neurodevelopmental differences have been linked to both executive function strengths and weaknesses, but evidence for associations between specific profiles of executive function and specific neurodevelopmental conditions is mixed. In this exploratory study, we adopt an unsupervised machine learning approach (self-organising maps), combined with k-means clustering to identify data-driven profiles of executive function in a transdiagnostic sample of 566 neurodivergent children aged 8-18 years old. We include measures designed to capture two distinct aspects of executive function: performance-based tasks designed to tap the state-like efficiency of cognitive skills under optimal conditions, and behaviour ratings suited to capturing the trait-like application of cognitive control in everyday contexts. Three profiles of executive function were identified: one had consistent difficulties across both types of assessments, while the other two had inconsistent profiles of predominantly rating- or predominantly task-based difficulties. Girls and children without a formal diagnosis were more likely to have an inconsistent profile of primarily task-based difficulties. Children with these different profiles had differences in academic achievement and mental health outcomes and could further be differentiated from a comparison group of children on both shared and profile-unique patterns of neural white matter organisation. Importantly, children's executive function profiles were not directly related to diagnostic categories or to dimensions of neurodiversity associated with specific diagnoses (e.g., hyperactivity, inattention, social communication). These findings support the idea that the two types of executive function assessments provide non-redundant information related to children's neurodevelopmental differences and that they should not be used interchangeably. The findings advance our understanding of executive function profiles and their relationship to behavioural outcomes and neural variation in neurodivergent populations.
执行功能是一个用于描述对思想、行动和情绪进行目标导向调节的统称，是神经多样性所涉及的一个重要维度，也是多种成人结局的可塑性预测指标。神经发育差异与执行功能的优势和劣势都有关联，但关于执行功能的特定特征与特定神经发育状况之间关联的证据并不一致。在这项探索性研究中，我们采用无监督机器学习方法（自组织映射），结合k均值聚类，在一个由566名8至18岁神经发育异常儿童组成的跨诊断样本中识别数据驱动的执行功能特征。我们纳入了旨在捕捉执行功能两个不同方面的测量方法：基于表现的任务，旨在挖掘最佳条件下认知技能的类似状态的效率；行为评分，适合捕捉日常情境中认知控制的类似特质的应用。识别出了三种执行功能特征：一种在两种评估类型中都存在持续困难，而另外两种的特征不一致，主要是基于评分或基于任务的困难。女孩和未接受正式诊断的儿童更有可能具有主要基于任务的困难的不一致特征。具有这些不同特征的儿童在学业成绩和心理健康结局方面存在差异，并且在神经白质组织的共享模式和特定特征模式上都可以与儿童对照组进一步区分开来。重要的是，儿童的执行功能特征与诊断类别或与特定诊断相关的神经多样性维度（例如多动、注意力不集中、社交沟通）没有直接关系。这些发现支持了这样一种观点，即两种执行功能评估提供了与儿童神经发育差异相关的非冗余信息，并且它们不应互换使用。这些发现推进了我们对执行功能特征及其与神经发育异常人群行为结局和神经变异之间关系的理解。
