Back to the main page.
Bug 872 - implement and test statfun_glm
| Status | CLOSED WONTFIX |
| Reported | 2011-08-17 10:50:00 +0200 |
| Modified | 2019-08-10 12:31:19 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P1 enhancement |
| Assigned to: | Stephen Whitmarsh |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Stephen Whitmarsh - 2011-08-17 10:50:37 +0200
- make example designs & contrasts - discuss design orthogonality - discuss T statistics, & testing using randomization inventorising cfg.function: - demean/ztransform data - keep mean (for ztransform) in case mean is explicitly modelled - demean/ztransform design - keep mean (for ztransform) in case mean is explicitly modelled - output designmatrix/design*design' - Consider implementing orthogonalization of design for including confound regressors.
Diego Lozano Soldevilla - 2013-02-13 13:39:42 +0100
could you give an update on this? If you have any relevant code, let us know. Otherwise we will think about having someone else implement it.
Robert Oostenveld - 2015-02-11 09:45:37 +0100
I don't think this is ever going to happen. GLMs and contrasts are just too flexible for us to provide an accessible level of documentation to go with it. Supporting the users with ft_statfun_glm therefore would be too much of a burden. If people want it, they can make their own.
Johanna - 2015-02-11 10:06:31 +0100
Created attachment 700 pilot version of ft_statfun_glm
Johanna - 2015-02-11 10:07:08 +0100
JM had made a version and shared it with me, and I used it happily. I modified it slightly. I've attached it here (for someone to use/find in future; hopefully that's ok with JM!!) but I understand if it seems too much work to support.