Dear Michael,
>Dear Rik,
>what is the theoretical and practical difference between
>the "effects of interest" F-contrast in the form
>[ 1 0 0, 0 1 0, 0 0 1] and a single F-contrast entered as
>[ 1 1 1 ] ?
The F-contrast spans the entire space spanned by the three vectors giben by
the corresponding regressors in your design matrix, and looks for variance
along any direction within this space.
The "single F contrast [1 1 1]" (which is really a t contrast) looks only
along a single line withinf that space.
To be specific, say that these three columns model three different
conditions and further say that for a given voxel conditions have no
effect, whereas the third has. The t contrast [1 1 1] tests for the average
effect across these three conditions, and is quite likely to be
insignificant due to the "dilution" by conditions 1 and 2. The F contrast
[1 0 0; 0 1 0; 0 0 1] simultaneously tests all possible t-contrasts within
this space, notably e.g. the contrast [0 0 1] which is likely to show a
large effect.
Hence, the t-contrast [1 1 1] tests for a much more restricted hypothesis
than does the F contrast [1 0 0; 0 1 0; 0 0 1].
Good luck Jesper
>Michael.
>--
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
>Dr. Michael Erb
>Sektion f. experimentelle Kernspinresonanz des ZNS
>Abteilung Neuroradiologie
>Universitaetsklinikum
>Hoppe-Seyler_Str. 3, D-72076 Tuebingen
>Tel.: +49(0)7071/2985398 priv. +49(0)7071/61559
>Fax.: +49(0)7071/294371
>e-mail: <[log in to unmask]>
>www: http://www.medizin.uni-tuebingen.de/nrad/sektion
><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>
>
Jesper Andersson
Wellcome Dept. of Cognitive Neurology
12 Queen Square
London WC1N 3BG
phone: 44 171 833 7484
fax: 44 171 813 1420
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|