Saturday, Jul 22: 1:00 PM - 5:00 PM
1365
Educational Course - Half Day (4 hours)
Palais
Room: 516AB
The evergreen popularity of neuroimaging has resulted in a wealth of findings, making it difficult for researchers to identify robust from irreproducible findings. Recent advances in meta-analysis have produced an impressive range of tools and methods to synthesize this vast literature into reliable and interpretable syntheses. Yet, researchers often lack guidance on how to best connect data and tools to perform meta-analyses in an open and reproducible manner. Participating researchers and clinicians will leave equipped with a decision chart to navigate their choices of data and tools in order to best conduct and interpret meta-analyses in line with their research goals.
1) Understand and apply the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) using a tutorial dataset. Participants will interpret inclusion criteria and make decisions about study eligibility.
2) Evaluate and differentiate the landscape of methods and tools used for neuroimaging meta-analyses. They should be able to read a new meta-analysis and critically consider its conclusions in light of the particular method's strengths and limitations, as well as other common sources of statistical error.
3) Explain how to use previous meta-analytic results in new studies and meta-analyses. For example, they will know how to generate regions of interest when planning a study and how to replicate and extend existing meta-analyses
This course will be useful for researchers and clinicians with any of these three goals:
1) To read and understand a meta-analysis’ methods, limitations, and interpretation; 2) To use past meta-analyses to form new hypotheses and interpret results in novel research; and 3) To perform neuroimaging meta-analyses in an open and reproducible manner.
Presentations
Timing
----------
20 min
Learning outcome
--------------------------
Understanding the role of meta-analyses in neuroimaging and the broad families of methods.
Points to cover
----------------------
Explain the limitations of single studies in neuroimaging, and the relation to the reproducibility crisis.
Explain the main uses of meta-analyses: synthesizing past results, developing new hypotheses, and interpreting novel results.
Outline the broad families of methods: coordinate- versus image-based methods, and manual versus automated methods.
Timing
----------
20 min
Learning outcome
--------------------------
Describe the strengths and limitations of how meta-analyses help with reproducibility
Points to cover
----------------------
Explain the reproducibility crisis in neuroimaging
Describe how meta-analyses can help with reproducibility
Explain how meta-analyses cannot solve all problems of reproducibility
Describe different sources of error in meta-analyses
Explain issues with original studies not corrected with meta-analyses
Timing
----------
20 min lecture
Learning outcome
--------------------------
Understand the relative strengths and weaknesses between manual and semi-automatic meta-analyses
Points to cover
----------------------
Explain the continued utility of manual meta-analyses, even with a wealth of high-quality, automated tools.
Describe some past manual meta-analyses that were important in the field and illustrate several different methods, including some well-done meta-analyses with inconclusive results.
Briefly outline the steps and options for manually selecting and annotating studies (including, for example, the role of NeuroSynth-Compose, metaCurious and NeuroVault).
Introduce the idea of "semi-automated" meta-analyses, leveraging tools like Neurosynth-Compose, NeuroVault, and metaCurious.
Describe in greater detail the NeuroSynth approach, which automates study selection and coordinate extraction. This results in less precise selection criteria and less accurate coordinate extraction, but the possibility to perform meta-analysis on a much larger number of studies.
Explain when to use automated meta-analyses and point out some common misinterpretations, such as invalid claims of reverse inference.
Explain why it’s important to share results images instead of just reporting peak coordinates.
Explain the limitations of conventional manual meta-analysis methods
Timing
----------
20 min lecture
Learning outcome
--------------------------
Understand the considerations when choosing which meta-analytic method to apply.
Points to cover
----------------------
Review the landscape of meta-analytic methods - from small number of studies to literature exploration, and additionally compare methods for image-based outcomes versus diagnostic outcomes
Explain the benefits and pitfalls of different methods
Illustrate a decision tree of meta-analytic methods
Describe the key points of consideration when choosing a meta-analytic method
Describe open source tools for literature exploration and article annotation
Presenter
Kendra Oudyk, MA, McGill University
Neurology and Neurosurgery
Montreal, Quebec
Canada
Timing
----------
15 min lecture
Learning outcome
--------------------------
Apply the practical steps of dataset curation and Annotation using NeuroSynth-Compose
Points to cover
----------------------
Craft an effective search on Pubmed and import the results into NeuroSynth-Compose
Explain the steps in a PRISMA workflow
Explain the process of determining study eligibility
Presenter
James Kent, PhD, University of Texas at Austin
Psychology
Austin, TX
United States
Timing
----------
15 min lecture
Learning outcome
--------------------------
Apply the practical steps of Meta-Analysis specification/execution using NeuroSynth-Compose and NiMARE
Points to cover
----------------------
How to specify and execute a meta-analysis
How to share and edit a meta-analysis
Interpret the results of the meta-analysis
Presenter
Yifan Yu, University of Oxford Oxford, Select...
United Kingdom