Mixed-effects models

The use of linear mixed-effect models is becoming increasingly popular in neuroimaging. These models extend the simple linear models by allowing both fixed and random effects, and are particularly useful for clustered and hierarchical data. These are the materials from a session held at the MRC Cognition and Brain Sciences Unit, Cambridge, UK, where Delia introduced the concept of linear mixed-effect models and discussed when they should be used. Roni then demonstrated how these are implemented with R. Alex reviewed some advanced topics, including estimation of parameters (family parameters and random effects) as well as the use of Bayesian statistics.

线性混合效应模型在神经影像学领域的应用日益广泛。此类模型通过允许固定效应与随机效应的同时存在，扩展了简单的线性模型，特别适用于簇状和层次数据。这些材料来源于在英国剑桥MRC认知与脑科学单元举办的一次研讨会，会上Delia介绍了线性混合效应模型的概念，并讨论了其适用的场景。随后，Roni通过R语言演示了这些模型的实现方法。Alex则回顾了一些高级主题，包括参数估计（家族参数和随机效应）以及贝叶斯统计学的应用。