Hi Hilary,
> Hi
>
> A brief overview of my experiment:
>
> I have tested peoples recognition memory for faces and scenes using
> a scale
> of (sure new) 1 - 6 (sure old).
>
> I now want to see if there are any regions that show a linear
> increase in
> activation to increases to recognition confidence regardless of
> accuracy for
> faces, for scenes and for both, i.e 1>2>3>4>5.
>
> I have done some reading around and just wanted to double check how
> I would
> model this in Feat.
>
> Firstly I would make an EV for each response + missed trails for
> faces and
> scenes so I would have 14 in total.
>
> EV1 = 1 for faces
> EV2 = 2 for faces
> EV3 = 3 for faces
> EV4 = 4 for faces
> EV5 = 5 for faces
> EV6 = 6 for faces
> EV7 = missed trials for faces
> EV8 = 1 for scenes
> EV9 = 2 for scenes
> EV10 = 3 for scenes
> EV11 = 4 for scenes
> EV12 = 5 for scenes
> EV13 = 6 for scenes
> EV14 = missed trials for scenes
>
> I plan to model each of these against my downtime baseline by doing to
> following?
>
> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12
> EV13 EV14
> 1 0 0 0 0 0 0 0 0
> 0 0
> 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0
> 0 0 0
> 0 0 1 0 0 0 0 0 0
> 0 0
> 0 0 0
> .. etc
>
looks fine
> to look at additions of these evs to baseline I would do the
> following? (so
> I could, for example, look at the combined activation for 1 for faces
> +scenes)..
>
> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12
> EV13 EV14
> 1 0 0 0 0 0 0 1 0
> 0 0
> 0 0 0
> 0 1 0 0 0 0 0 0 1
> 0 0
> 0 0 0
> 0 0 1 0 0 0 0 0 0
> 1 0
> 0 0 0
> .. etc
>
more specifically this will test the mean activation across 1 for
faces and scenes.
> to look at a linear trend across confidence 1-5 for faces and scenes
> seperately I would do the following?
>
> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12
> EV13 EV14
> -2 -1 0 1 2 0 0 0 0 0 0
> 0 0 0
> 0 1 0 0 0 0 0 -2 -1 0 1
> 2 0 0
yes - although I am guessing that the 1 in the second contrast for EV2
is a typo.
> to look at a linear trend across confidence 1-5 for faces and scenes
> combined I would do the following?
>
> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8 EV9 EV10 EV11 EV12
> EV13 EV14
> -2 -1 0 1 2 0 0 -2 -1 0 1
> 2 0 0
>
>
> The main reason I am confused is because on the FSL help pages it
> gives the
> following example
>
> [-1 1 0] shows where the response to level 2 is greater than that
> for level
> 1. [-1 0 1] shows the general linear increase across all three levels.
>
> But doesn't the latter give you the actiavtion of C-A? How does FSL
> know
> that you want to look at the linear increase and not just subtract
> one EV
> from another?
Take a look at this thread for a previous discussion on this:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind06&L=FSL&P=R246087
Cheers, Mark.
|