2021 – Using Causal Methods to Evaluate the Role of Biomarkers

Schedule 2021

The conference on modern psychometric and statistical methods in cognitive research will be a virtual conference in 2021. The 2020 conference did not occur due to COVID-19 and was postponed to 2021. The theme will be “Using Causal Methods to Evaluate the Role of Biomarkers as Mechanisms of Alzheimer’s Disease and Dementia”. The growing availability of biomarkers of vascular and amyloid pathway mechanisms for dementia raise a number of methodologic challenges. Workgroups will focus on two types of methods: statistical tools that allow us to generalize from highly selected, specialized samples often included in biomarker studies to more typical population samples; and contemporary methods for evaluation of mediation. The biomarker studies that have generated much of the knowledge base about amyloid and vascular mechanisms for cognitive decline and dementia have typically utilized highly selected samples, and this raises a fundamental question about whether these findings apply to broader populations. Understanding how sampling bias affects substantive knowledge about processes and mechanisms of cognitive decline is critical for developing an empirical knowledge base about late life cognitive decline that encompasses the rapidly growing and increasingly diverse older population. Recent developments in the causal inference literature provide methods for studying effects of sampling bias and potentially adjusting empirical results to account for such biases. A set of methods are available to help transport associations between data sets and this is an area of recent progress in the causal inference literature. Biomarker studies are also important for studying biological mechanisms that link measured risk factors to cognitive outcomes. Recent developments in mediation and effect decomposition models are highly relevant to biomarker studies. Workgroups focused on mediation will evaluate the role of measured biomarkers in accounting for the effects of established social, biological, or behavioral risk factors on clinical outcomes.