How to avoid mismodelling in GLM-based fMRI data analysis: cross-validated Bayesian model selection.

Joram Soch, John-Dylan Haynes, Carsten Allefeld

November 2016 Neuroimage

Synopsis of Social media discussions

Discussions highlight appreciation for the article's novel approach, such as mentions of cross-validated Bayesian model selection, with some posts emphasizing its potential to improve analysis reliability. The tone is generally positive and focused on the technical merits, with examples like 'summarize one of your papers' and references to 'new research' that reflect both engagement and recognition of the publication's value.

A
Agreement
 Moderate agreement

Most posts acknowledge the importance of the publication, with some explicitly supporting its methods and findings.

I
Interest
 Moderate level of interest

Posts indicate a moderate level of curiosity, often referencing the methodology and potential benefits.

E
Engagement
 Moderate level of engagement

Discussion includes references to summarization efforts and the practical implications of the research, reflecting active engagement.

I
Impact
 Moderate level of impact

Users recognize the work's significance for improving fMRI analysis and model selection, suggesting notable influence.

Social Mentions

YouTube

1 Videos

Facebook

2 Posts

Twitter

3 Posts

Metrics

Video Views

90

Total Likes

3

Extended Reach

2,535

Social Features

6

Timeline: Posts about article

Top Social Media Posts

Posts referencing the article

Avoiding Mismodeling in fMRI Analysis Using Cross-Validated Bayesian Model Selection

Avoiding Mismodeling in fMRI Analysis Using Cross-Validated Bayesian Model Selection

This video discusses how to improve fMRI data analysis with General Linear Models by using cross-validated Bayesian model selection to choose the best model, enhancing reliability and reproducibility in neuroimaging studies.

Joram Soch

July 18, 2022

90 views

  • The Book of Statistical Proofs
    @StatProofBook (Twitter)

    RT @JoramSoch: I'm starting #PaperVideos: Summarize one of your papers in 5 min or less. Here's an example for our work on Bayesian model s…
    view full post

    May 2, 2023

    1

  • Joram Soch
    @JoramSoch (Twitter)

    I'm starting #PaperVideos: Summarize one of your papers in 5 min or less. Here's an example for our work on Bayesian model selection for general linear models in fMRI data analysis (sorry, it's 5:22 min
    view full post

    May 1, 2023

    3

    1

  • City Research Online
    @City_Research (Twitter)

    #openaccess New research in CRO: September 23, 2019 at 10:57AM How to avoid mismodelling in GLM-based fMRI data analysis: cross-validated Bayesian model selection https://t.co/dETefG77lb
    view full post

    September 23, 2019

Abstract Synopsis

  • The article discusses the use of General Linear Models (GLMs) in analyzing fMRI data, highlighting their flexibility but also the challenges in choosing the right model to avoid underfitting or overfitting.
  • It introduces a new method called crossvalidated Bayesian model selection (cvBMS), which helps determine the best GLM for different brain areas and data sets, even when models are not directly comparable.
  • The approach is validated through simulations and real data, aiming to improve the reliability and reproducibility of fMRI research by systematically comparing multiple models.]