Determining significance in neuroimaging studies using covariate-modulated false discovery rate

Posted in Science on September 17, 2008


Determining significance in neuroimaging studies using covariate-modulated false discovery rate

Lecture slides:

  • Advances in False Discovery Rate control applied in Neuroimaging
  • Outline
  • Voxel-based morphometry
  • Vertex-based morphometry
  • Pial (outside) surface
  • White matter surface
  • Wireframe
  • Multiple comparisons
  • Searching for interesting results
  • Outline
  • Definition
  • Markov Random Fields
  • Permutation tests.
  • Weakness with blob-based methods
  • The original concept
  • The method
  • Graphical
  • FDR in practice
  • Outline
  • Local FDR
  • The model
  • Definition
  • Why local?
  • Histogram of Z-scores
  • Requirements
  • Calculating the local FDR
  • The mixture f (z)
  • The numerator p0f0(z)
  • Outline
  • Covariate Modulated FDR
  • Microarray example
  • The model
  • Definition
  • Compare to Local FDR
  • The model, II
  • Calculating cmFDR
  • Observation
  • Bin the covariates
  • Hyperpriors
  • posterior density
  • Approximate
  • cmFDR with one bin
  • Outline
  • Genetic variation and cortical thickness
  • BDNF -663
  • cmFDR reasoning
  • Comparison
  • Bibliography

Author: Glenn Lawyer, University Of Oslo

Watch Video

Tags: Science, Lectures, Statistics, Computer Science, Machine Learning, VideoLectures.Net, Mathematics