Type I and type II errors for Multiple Simultaneous Hypothesis Testing

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

author:Glenn Lawyer, Faculty of Medicine, University of Oslo
published: May 23, 2007, recorded: May 2007, views: 172

Slides

Slides
0:02	Advances in False Discovery Rate control applied in Neuroimaging
0:29	Outline
1:23	Outline
1:31	Voxel-based morphometry
2:09	Voxel-based morphometry
2:18	Vertex-based morphometry
2:54	Pial (outside) surface
3:03	White matter surface
3:11	Wireframe
3:57	Multiple comparisons
4:04	Multiple comparisons
4:27	Searching for interesting results
5:06	Outline
5:58	Definition
6:06	Markov Random Fields
6:29	Markov Random Fields
6:57	Permutation tests.
7:25	Weakness with blob-based methods
8:03	The original concept
8:07	The method
8:11	Graphical
9:03	FDR in practice
9:25	Outline
9:34	Local FDR
9:44	The model
10:38	The model, cont.
11:02	Definition
11:11	Definition
11:30	Why local?
11:59	Histogram of Z-scores
13:09	Requirements
13:34	Calculating the local FDR
14:06	The mixture f (z)
14:12	The mixture f (z)
14:24	The mixture f (z)
14:58	The numerator p0f0(z)
15:24	The numerator p0f0(z)
15:52	The numerator p0f0(z)
16:49	Outline
17:03	Covariate Modulated FDR
17:38	Microarray example
18:48	The model
19:00	The model
19:08	Definition
19:22	Definition
19:35	Definition
20:17	Compare to Local FDR
20:42	The model, II
21:39	Calculating cmFDR
21:46	Observation
22:39	Bin the covariates
23:40	Hyperpriors
24:01	posterior density
24:36	posterior density, cont.
24:45	Approximate
25:11	Approximate
25:35	cmFDR with one bin
26:54	Outline
27:00	Genetic variation and cortical thickness
28:32	BDNF -663
30:55	cmFDR reasoning
31:19	Comparison
31:53	Comparison
35:18	Bibliography

Related content

Visitors who watched this lecture also watched...